{"base_code":{"magic":"pir","trainable":true,"version":1},"program":{"regions":[{"#":"region_0","blocks":[{"#":"block_0","args":[],"ops":[{"#":"p","A":[0,1,1,"linear_73.b_0"],"DA":[],"O":{"%":1,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[21],"NCHW",[],0]}},"OA":[1,0,1],"QA":[]},{"#":"p","A":[0,1,1,"linear_73.w_0"],"DA":[],"O":{"%":2,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[768,21],"NCHW",[],0]}},"OA":[1,0,1],"QA":[]},{"#":"p","A":[0,1,1,"linear_72.b_0"],"DA":[],"O":{"%":3,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[768],"NCHW",[],0]}},"OA":[1,0,1],"QA":[]},{"#":"p","A":[0,1,1,"linear_72.w_0"],"DA":[],"O":{"%":4,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[768,768],"NCHW",[],0]}},"OA":[1,0,1],"QA":[]},{"#":"p","A":[0,1,1,"layer_norm_24.b_0"],"DA":[],"O":{"%":5,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[768],"NCHW",[],0]}},"OA":[1,0,1],"QA":[]},{"#":"p","A":[0,1,1,"layer_norm_24.w_0"],"DA":[],"O":{"%":6,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[768],"NCHW",[],0]}},"OA":[1,0,1],"QA":[]},{"#":"p","A":[0,1,1,"layer_norm_23.b_0"],"DA":[],"O":{"%":7,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[768],"NCHW",[],0]}},"OA":[1,0,1],"QA":[]},{"#":"p","A":[0,1,1,"layer_norm_23.w_0"],"DA":[],"O":{"%":8,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[768],"NCHW",[],0]}},"OA":[1,0,1],"QA":[]},{"#":"p","A":[0,1,1,"linear_71.b_0"],"DA":[],"O":{"%":9,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[768],"NCHW",[],0]}},"OA":[1,0,1],"QA":[]},{"#":"p","A":[0,1,1,"linear_71.w_0"],"DA":[],"O":{"%":10,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[3072,768],"NCHW",[],0]}},"OA":[1,0,1],"QA":[]},{"#":"p","A":[0,1,1,"linear_70.b_0"],"DA":[],"O":{"%":11,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[3072],"NCHW",[],0]}},"OA":[1,0,1],"QA":[]},{"#":"p","A":[0,1,1,"linear_70.w_0"],"DA":[],"O":{"%":12,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[768,3072],"NCHW",[],0]}},"OA":[1,0,1],"QA":[]},{"#":"p","A":[0,1,1,"linear_69.b_0"],"DA":[],"O":{"%":13,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[768],"NCHW",[],0]}},"OA":[1,0,1],"QA":[]},{"#":"p","A":[0,1,1,"linear_69.w_0"],"DA":[],"O":{"%":14,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[768,768],"NCHW",[],0]}},"OA":[1,0,1],"QA":[]},{"#":"p","A":[0,1,1,"linear_68.b_0"],"DA":[],"O":{"%":15,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[768],"NCHW",[],0]}},"OA":[1,0,1],"QA":[]},{"#":"p","A":[0,1,1,"linear_68.w_0"],"DA":[],"O":{"%":16,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[768,768],"NCHW",[],0]}},"OA":[1,0,1],"QA":[]},{"#":"p","A":[0,1,1,"linear_67.b_0"],"DA":[],"O":{"%":17,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[768],"NCHW",[],0]}},"OA":[1,0,1],"QA":[]},{"#":"p","A":[0,1,1,"linear_67.w_0"],"DA":[],"O":{"%":18,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[768,768],"NCHW",[],0]}},"OA":[1,0,1],"QA":[]},{"#":"p","A":[0,1,1,"linear_66.b_0"],"DA":[],"O":{"%":19,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[768],"NCHW",[],0]}},"OA":[1,0,1],"QA":[]},{"#":"p","A":[0,1,1,"linear_66.w_0"],"DA":[],"O":{"%":20,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[768,768],"NCHW",[],0]}},"OA":[1,0,1],"QA":[]},{"#":"p","A":[0,1,1,"layer_norm_22.b_0"],"DA":[],"O":{"%":21,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[768],"NCHW",[],0]}},"OA":[1,0,1],"QA":[]},{"#":"p","A":[0,1,1,"layer_norm_22.w_0"],"DA":[],"O":{"%":22,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[768],"NCHW",[],0]}},"OA":[1,0,1],"QA":[]},{"#":"p","A":[0,1,1,"layer_norm_21.b_0"],"DA":[],"O":{"%":23,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[768],"NCHW",[],0]}},"OA":[1,0,1],"QA":[]},{"#":"p","A":[0,1,1,"layer_norm_21.w_0"],"DA":[],"O":{"%":24,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[768],"NCHW",[],0]}},"OA":[1,0,1],"QA":[]},{"#":"p","A":[0,1,1,"linear_65.b_0"],"DA":[],"O":{"%":25,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[768],"NCHW",[],0]}},"OA":[1,0,1],"QA":[]},{"#":"p","A":[0,1,1,"linear_65.w_0"],"DA":[],"O":{"%":26,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[3072,768],"NCHW",[],0]}},"OA":[1,0,1],"QA":[]},{"#":"p","A":[0,1,1,"linear_64.b_0"],"DA":[],"O":{"%":27,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[3072],"NCHW",[],0]}},"OA":[1,0,1],"QA":[]},{"#":"p","A":[0,1,1,"linear_64.w_0"],"DA":[],"O":{"%":28,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[768,3072],"NCHW",[],0]}},"OA":[1,0,1],"QA":[]},{"#":"p","A":[0,1,1,"linear_63.b_0"],"DA":[],"O":{"%":29,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[768],"NCHW",[],0]}},"OA":[1,0,1],"QA":[]},{"#":"p","A":[0,1,1,"linear_63.w_0"],"DA":[],"O":{"%":30,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[768,768],"NCHW",[],0]}},"OA":[1,0,1],"QA":[]},{"#":"p","A":[0,1,1,"linear_62.b_0"],"DA":[],"O":{"%":31,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[768],"NCHW",[],0]}},"OA":[1,0,1],"QA":[]},{"#":"p","A":[0,1,1,"linear_62.w_0"],"DA":[],"O":{"%":32,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[768,768],"NCHW",[],0]}},"OA":[1,0,1],"QA":[]},{"#":"p","A":[0,1,1,"linear_61.b_0"],"DA":[],"O":{"%":33,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[768],"NCHW",[],0]}},"OA":[1,0,1],"QA":[]},{"#":"p","A":[0,1,1,"linear_61.w_0"],"DA":[],"O":{"%":34,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[768,768],"NCHW",[],0]}},"OA":[1,0,1],"QA":[]},{"#":"p","A":[0,1,1,"linear_60.b_0"],"DA":[],"O":{"%":35,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[768],"NCHW",[],0]}},"OA":[1,0,1],"QA":[]},{"#":"p","A":[0,1,1,"linear_60.w_0"],"DA":[],"O":{"%":36,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[768,768],"NCHW",[],0]}},"OA":[1,0,1],"QA":[]},{"#":"p","A":[0,1,1,"layer_norm_20.b_0"],"DA":[],"O":{"%":37,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[768],"NCHW",[],0]}},"OA":[1,0,1],"QA":[]},{"#":"p","A":[0,1,1,"layer_norm_20.w_0"],"DA":[],"O":{"%":38,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[768],"NCHW",[],0]}},"OA":[1,0,1],"QA":[]},{"#":"p","A":[0,1,1,"layer_norm_19.b_0"],"DA":[],"O":{"%":39,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[768],"NCHW",[],0]}},"OA":[1,0,1],"QA":[]},{"#":"p","A":[0,1,1,"layer_norm_19.w_0"],"DA":[],"O":{"%":40,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[768],"NCHW",[],0]}},"OA":[1,0,1],"QA":[]},{"#":"p","A":[0,1,1,"linear_59.b_0"],"DA":[],"O":{"%":41,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[768],"NCHW",[],0]}},"OA":[1,0,1],"QA":[]},{"#":"p","A":[0,1,1,"linear_59.w_0"],"DA":[],"O":{"%":42,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[3072,768],"NCHW",[],0]}},"OA":[1,0,1],"QA":[]},{"#":"p","A":[0,1,1,"linear_58.b_0"],"DA":[],"O":{"%":43,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[3072],"NCHW",[],0]}},"OA":[1,0,1],"QA":[]},{"#":"p","A":[0,1,1,"linear_58.w_0"],"DA":[],"O":{"%":44,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[768,3072],"NCHW",[],0]}},"OA":[1,0,1],"QA":[]},{"#":"p","A":[0,1,1,"linear_57.b_0"],"DA":[],"O":{"%":45,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[768],"NCHW",[],0]}},"OA":[1,0,1],"QA":[]},{"#":"p","A":[0,1,1,"linear_57.w_0"],"DA":[],"O":{"%":46,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[768,768],"NCHW",[],0]}},"OA":[1,0,1],"QA":[]},{"#":"p","A":[0,1,1,"linear_56.b_0"],"DA":[],"O":{"%":47,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[768],"NCHW",[],0]}},"OA":[1,0,1],"QA":[]},{"#":"p","A":[0,1,1,"linear_56.w_0"],"DA":[],"O":{"%":48,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[768,768],"NCHW",[],0]}},"OA":[1,0,1],"QA":[]},{"#":"p","A":[0,1,1,"linear_55.b_0"],"DA":[],"O":{"%":49,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[768],"NCHW",[],0]}},"OA":[1,0,1],"QA":[]},{"#":"p","A":[0,1,1,"linear_55.w_0"],"DA":[],"O":{"%":50,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[768,768],"NCHW",[],0]}},"OA":[1,0,1],"QA":[]},{"#":"p","A":[0,1,1,"linear_54.b_0"],"DA":[],"O":{"%":51,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[768],"NCHW",[],0]}},"OA":[1,0,1],"QA":[]},{"#":"p","A":[0,1,1,"linear_54.w_0"],"DA":[],"O":{"%":52,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[768,768],"NCHW",[],0]}},"OA":[1,0,1],"QA":[]},{"#":"p","A":[0,1,1,"layer_norm_18.b_0"],"DA":[],"O":{"%":53,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[768],"NCHW",[],0]}},"OA":[1,0,1],"QA":[]},{"#":"p","A":[0,1,1,"layer_norm_18.w_0"],"DA":[],"O":{"%":54,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[768],"NCHW",[],0]}},"OA":[1,0,1],"QA":[]},{"#":"p","A":[0,1,1,"layer_norm_17.b_0"],"DA":[],"O":{"%":55,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[768],"NCHW",[],0]}},"OA":[1,0,1],"QA":[]},{"#":"p","A":[0,1,1,"layer_norm_17.w_0"],"DA":[],"O":{"%":56,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[768],"NCHW",[],0]}},"OA":[1,0,1],"QA":[]},{"#":"p","A":[0,1,1,"linear_53.b_0"],"DA":[],"O":{"%":57,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[768],"NCHW",[],0]}},"OA":[1,0,1],"QA":[]},{"#":"p","A":[0,1,1,"linear_53.w_0"],"DA":[],"O":{"%":58,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[3072,768],"NCHW",[],0]}},"OA":[1,0,1],"QA":[]},{"#":"p","A":[0,1,1,"linear_52.b_0"],"DA":[],"O":{"%":59,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[3072],"NCHW",[],0]}},"OA":[1,0,1],"QA":[]},{"#":"p","A":[0,1,1,"linear_52.w_0"],"DA":[],"O":{"%":60,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[768,3072],"NCHW",[],0]}},"OA":[1,0,1],"QA":[]},{"#":"p","A":[0,1,1,"linear_51.b_0"],"DA":[],"O":{"%":61,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[768],"NCHW",[],0]}},"OA":[1,0,1],"QA":[]},{"#":"p","A":[0,1,1,"linear_51.w_0"],"DA":[],"O":{"%":62,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[768,768],"NCHW",[],0]}},"OA":[1,0,1],"QA":[]},{"#":"p","A":[0,1,1,"linear_50.b_0"],"DA":[],"O":{"%":63,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[768],"NCHW",[],0]}},"OA":[1,0,1],"QA":[]},{"#":"p","A":[0,1,1,"linear_50.w_0"],"DA":[],"O":{"%":64,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[768,768],"NCHW",[],0]}},"OA":[1,0,1],"QA":[]},{"#":"p","A":[0,1,1,"linear_49.b_0"],"DA":[],"O":{"%":65,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[768],"NCHW",[],0]}},"OA":[1,0,1],"QA":[]},{"#":"p","A":[0,1,1,"linear_49.w_0"],"DA":[],"O":{"%":66,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[768,768],"NCHW",[],0]}},"OA":[1,0,1],"QA":[]},{"#":"p","A":[0,1,1,"linear_48.b_0"],"DA":[],"O":{"%":67,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[768],"NCHW",[],0]}},"OA":[1,0,1],"QA":[]},{"#":"p","A":[0,1,1,"linear_48.w_0"],"DA":[],"O":{"%":68,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[768,768],"NCHW",[],0]}},"OA":[1,0,1],"QA":[]},{"#":"p","A":[0,1,1,"layer_norm_16.b_0"],"DA":[],"O":{"%":69,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[768],"NCHW",[],0]}},"OA":[1,0,1],"QA":[]},{"#":"p","A":[0,1,1,"layer_norm_16.w_0"],"DA":[],"O":{"%":70,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[768],"NCHW",[],0]}},"OA":[1,0,1],"QA":[]},{"#":"p","A":[0,1,1,"layer_norm_15.b_0"],"DA":[],"O":{"%":71,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[768],"NCHW",[],0]}},"OA":[1,0,1],"QA":[]},{"#":"p","A":[0,1,1,"layer_norm_15.w_0"],"DA":[],"O":{"%":72,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[768],"NCHW",[],0]}},"OA":[1,0,1],"QA":[]},{"#":"p","A":[0,1,1,"linear_47.b_0"],"DA":[],"O":{"%":73,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[768],"NCHW",[],0]}},"OA":[1,0,1],"QA":[]},{"#":"p","A":[0,1,1,"linear_47.w_0"],"DA":[],"O":{"%":74,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[3072,768],"NCHW",[],0]}},"OA":[1,0,1],"QA":[]},{"#":"p","A":[0,1,1,"linear_46.b_0"],"DA":[],"O":{"%":75,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[3072],"NCHW",[],0]}},"OA":[1,0,1],"QA":[]},{"#":"p","A":[0,1,1,"linear_46.w_0"],"DA":[],"O":{"%":76,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[768,3072],"NCHW",[],0]}},"OA":[1,0,1],"QA":[]},{"#":"p","A":[0,1,1,"linear_45.b_0"],"DA":[],"O":{"%":77,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[768],"NCHW",[],0]}},"OA":[1,0,1],"QA":[]},{"#":"p","A":[0,1,1,"linear_45.w_0"],"DA":[],"O":{"%":78,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[768,768],"NCHW",[],0]}},"OA":[1,0,1],"QA":[]},{"#":"p","A":[0,1,1,"linear_44.b_0"],"DA":[],"O":{"%":79,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[768],"NCHW",[],0]}},"OA":[1,0,1],"QA":[]},{"#":"p","A":[0,1,1,"linear_44.w_0"],"DA":[],"O":{"%":80,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[768,768],"NCHW",[],0]}},"OA":[1,0,1],"QA":[]},{"#":"p","A":[0,1,1,"linear_43.b_0"],"DA":[],"O":{"%":81,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[768],"NCHW",[],0]}},"OA":[1,0,1],"QA":[]},{"#":"p","A":[0,1,1,"linear_43.w_0"],"DA":[],"O":{"%":82,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[768,768],"NCHW",[],0]}},"OA":[1,0,1],"QA":[]},{"#":"p","A":[0,1,1,"linear_42.b_0"],"DA":[],"O":{"%":83,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[768],"NCHW",[],0]}},"OA":[1,0,1],"QA":[]},{"#":"p","A":[0,1,1,"linear_42.w_0"],"DA":[],"O":{"%":84,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[768,768],"NCHW",[],0]}},"OA":[1,0,1],"QA":[]},{"#":"p","A":[0,1,1,"layer_norm_14.b_0"],"DA":[],"O":{"%":85,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[768],"NCHW",[],0]}},"OA":[1,0,1],"QA":[]},{"#":"p","A":[0,1,1,"layer_norm_14.w_0"],"DA":[],"O":{"%":86,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[768],"NCHW",[],0]}},"OA":[1,0,1],"QA":[]},{"#":"p","A":[0,1,1,"layer_norm_13.b_0"],"DA":[],"O":{"%":87,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[768],"NCHW",[],0]}},"OA":[1,0,1],"QA":[]},{"#":"p","A":[0,1,1,"layer_norm_13.w_0"],"DA":[],"O":{"%":88,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[768],"NCHW",[],0]}},"OA":[1,0,1],"QA":[]},{"#":"p","A":[0,1,1,"linear_41.b_0"],"DA":[],"O":{"%":89,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[768],"NCHW",[],0]}},"OA":[1,0,1],"QA":[]},{"#":"p","A":[0,1,1,"linear_41.w_0"],"DA":[],"O":{"%":90,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[3072,768],"NCHW",[],0]}},"OA":[1,0,1],"QA":[]},{"#":"p","A":[0,1,1,"linear_40.b_0"],"DA":[],"O":{"%":91,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[3072],"NCHW",[],0]}},"OA":[1,0,1],"QA":[]},{"#":"p","A":[0,1,1,"linear_40.w_0"],"DA":[],"O":{"%":92,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[768,3072],"NCHW",[],0]}},"OA":[1,0,1],"QA":[]},{"#":"p","A":[0,1,1,"linear_39.b_0"],"DA":[],"O":{"%":93,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[768],"NCHW",[],0]}},"OA":[1,0,1],"QA":[]},{"#":"p","A":[0,1,1,"linear_39.w_0"],"DA":[],"O":{"%":94,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[768,768],"NCHW",[],0]}},"OA":[1,0,1],"QA":[]},{"#":"p","A":[0,1,1,"linear_38.b_0"],"DA":[],"O":{"%":95,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[768],"NCHW",[],0]}},"OA":[1,0,1],"QA":[]},{"#":"p","A":[0,1,1,"linear_38.w_0"],"DA":[],"O":{"%":96,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[768,768],"NCHW",[],0]}},"OA":[1,0,1],"QA":[]},{"#":"p","A":[0,1,1,"linear_37.b_0"],"DA":[],"O":{"%":97,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[768],"NCHW",[],0]}},"OA":[1,0,1],"QA":[]},{"#":"p","A":[0,1,1,"linear_37.w_0"],"DA":[],"O":{"%":98,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[768,768],"NCHW",[],0]}},"OA":[1,0,1],"QA":[]},{"#":"p","A":[0,1,1,"linear_36.b_0"],"DA":[],"O":{"%":99,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[768],"NCHW",[],0]}},"OA":[1,0,1],"QA":[]},{"#":"p","A":[0,1,1,"linear_36.w_0"],"DA":[],"O":{"%":100,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[768,768],"NCHW",[],0]}},"OA":[1,0,1],"QA":[]},{"#":"p","A":[0,1,1,"layer_norm_12.b_0"],"DA":[],"O":{"%":101,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[768],"NCHW",[],0]}},"OA":[1,0,1],"QA":[]},{"#":"p","A":[0,1,1,"layer_norm_12.w_0"],"DA":[],"O":{"%":102,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[768],"NCHW",[],0]}},"OA":[1,0,1],"QA":[]},{"#":"p","A":[0,1,1,"layer_norm_11.b_0"],"DA":[],"O":{"%":103,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[768],"NCHW",[],0]}},"OA":[1,0,1],"QA":[]},{"#":"p","A":[0,1,1,"layer_norm_11.w_0"],"DA":[],"O":{"%":104,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[768],"NCHW",[],0]}},"OA":[1,0,1],"QA":[]},{"#":"p","A":[0,1,1,"linear_35.b_0"],"DA":[],"O":{"%":105,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[768],"NCHW",[],0]}},"OA":[1,0,1],"QA":[]},{"#":"p","A":[0,1,1,"linear_35.w_0"],"DA":[],"O":{"%":106,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[3072,768],"NCHW",[],0]}},"OA":[1,0,1],"QA":[]},{"#":"p","A":[0,1,1,"linear_34.b_0"],"DA":[],"O":{"%":107,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[3072],"NCHW",[],0]}},"OA":[1,0,1],"QA":[]},{"#":"p","A":[0,1,1,"linear_34.w_0"],"DA":[],"O":{"%":108,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[768,3072],"NCHW",[],0]}},"OA":[1,0,1],"QA":[]},{"#":"p","A":[0,1,1,"linear_33.b_0"],"DA":[],"O":{"%":109,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[768],"NCHW",[],0]}},"OA":[1,0,1],"QA":[]},{"#":"p","A":[0,1,1,"linear_33.w_0"],"DA":[],"O":{"%":110,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[768,768],"NCHW",[],0]}},"OA":[1,0,1],"QA":[]},{"#":"p","A":[0,1,1,"linear_32.b_0"],"DA":[],"O":{"%":111,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[768],"NCHW",[],0]}},"OA":[1,0,1],"QA":[]},{"#":"p","A":[0,1,1,"linear_32.w_0"],"DA":[],"O":{"%":112,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[768,768],"NCHW",[],0]}},"OA":[1,0,1],"QA":[]},{"#":"p","A":[0,1,1,"linear_31.b_0"],"DA":[],"O":{"%":113,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[768],"NCHW",[],0]}},"OA":[1,0,1],"QA":[]},{"#":"p","A":[0,1,1,"linear_31.w_0"],"DA":[],"O":{"%":114,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[768,768],"NCHW",[],0]}},"OA":[1,0,1],"QA":[]},{"#":"p","A":[0,1,1,"linear_30.b_0"],"DA":[],"O":{"%":115,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[768],"NCHW",[],0]}},"OA":[1,0,1],"QA":[]},{"#":"p","A":[0,1,1,"linear_30.w_0"],"DA":[],"O":{"%":116,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[768,768],"NCHW",[],0]}},"OA":[1,0,1],"QA":[]},{"#":"p","A":[0,1,1,"layer_norm_10.b_0"],"DA":[],"O":{"%":117,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[768],"NCHW",[],0]}},"OA":[1,0,1],"QA":[]},{"#":"p","A":[0,1,1,"layer_norm_10.w_0"],"DA":[],"O":{"%":118,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[768],"NCHW",[],0]}},"OA":[1,0,1],"QA":[]},{"#":"p","A":[0,1,1,"layer_norm_9.b_0"],"DA":[],"O":{"%":119,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[768],"NCHW",[],0]}},"OA":[1,0,1],"QA":[]},{"#":"p","A":[0,1,1,"layer_norm_9.w_0"],"DA":[],"O":{"%":120,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[768],"NCHW",[],0]}},"OA":[1,0,1],"QA":[]},{"#":"p","A":[0,1,1,"linear_29.b_0"],"DA":[],"O":{"%":121,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[768],"NCHW",[],0]}},"OA":[1,0,1],"QA":[]},{"#":"p","A":[0,1,1,"linear_29.w_0"],"DA":[],"O":{"%":122,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[3072,768],"NCHW",[],0]}},"OA":[1,0,1],"QA":[]},{"#":"p","A":[0,1,1,"linear_28.b_0"],"DA":[],"O":{"%":123,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[3072],"NCHW",[],0]}},"OA":[1,0,1],"QA":[]},{"#":"p","A":[0,1,1,"linear_28.w_0"],"DA":[],"O":{"%":124,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[768,3072],"NCHW",[],0]}},"OA":[1,0,1],"QA":[]},{"#":"p","A":[0,1,1,"linear_27.b_0"],"DA":[],"O":{"%":125,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[768],"NCHW",[],0]}},"OA":[1,0,1],"QA":[]},{"#":"p","A":[0,1,1,"linear_27.w_0"],"DA":[],"O":{"%":126,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[768,768],"NCHW",[],0]}},"OA":[1,0,1],"QA":[]},{"#":"p","A":[0,1,1,"linear_26.b_0"],"DA":[],"O":{"%":127,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[768],"NCHW",[],0]}},"OA":[1,0,1],"QA":[]},{"#":"p","A":[0,1,1,"linear_26.w_0"],"DA":[],"O":{"%":128,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[768,768],"NCHW",[],0]}},"OA":[1,0,1],"QA":[]},{"#":"p","A":[0,1,1,"linear_25.b_0"],"DA":[],"O":{"%":129,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[768],"NCHW",[],0]}},"OA":[1,0,1],"QA":[]},{"#":"p","A":[0,1,1,"linear_25.w_0"],"DA":[],"O":{"%":130,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[768,768],"NCHW",[],0]}},"OA":[1,0,1],"QA":[]},{"#":"p","A":[0,1,1,"linear_24.b_0"],"DA":[],"O":{"%":131,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[768],"NCHW",[],0]}},"OA":[1,0,1],"QA":[]},{"#":"p","A":[0,1,1,"linear_24.w_0"],"DA":[],"O":{"%":132,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[768,768],"NCHW",[],0]}},"OA":[1,0,1],"QA":[]},{"#":"p","A":[0,1,1,"layer_norm_8.b_0"],"DA":[],"O":{"%":133,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[768],"NCHW",[],0]}},"OA":[1,0,1],"QA":[]},{"#":"p","A":[0,1,1,"layer_norm_8.w_0"],"DA":[],"O":{"%":134,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[768],"NCHW",[],0]}},"OA":[1,0,1],"QA":[]},{"#":"p","A":[0,1,1,"layer_norm_7.b_0"],"DA":[],"O":{"%":135,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[768],"NCHW",[],0]}},"OA":[1,0,1],"QA":[]},{"#":"p","A":[0,1,1,"layer_norm_7.w_0"],"DA":[],"O":{"%":136,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[768],"NCHW",[],0]}},"OA":[1,0,1],"QA":[]},{"#":"p","A":[0,1,1,"linear_23.b_0"],"DA":[],"O":{"%":137,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[768],"NCHW",[],0]}},"OA":[1,0,1],"QA":[]},{"#":"p","A":[0,1,1,"linear_23.w_0"],"DA":[],"O":{"%":138,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[3072,768],"NCHW",[],0]}},"OA":[1,0,1],"QA":[]},{"#":"p","A":[0,1,1,"linear_22.b_0"],"DA":[],"O":{"%":139,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[3072],"NCHW",[],0]}},"OA":[1,0,1],"QA":[]},{"#":"p","A":[0,1,1,"linear_22.w_0"],"DA":[],"O":{"%":140,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[768,3072],"NCHW",[],0]}},"OA":[1,0,1],"QA":[]},{"#":"p","A":[0,1,1,"linear_21.b_0"],"DA":[],"O":{"%":141,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[768],"NCHW",[],0]}},"OA":[1,0,1],"QA":[]},{"#":"p","A":[0,1,1,"linear_21.w_0"],"DA":[],"O":{"%":142,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[768,768],"NCHW",[],0]}},"OA":[1,0,1],"QA":[]},{"#":"p","A":[0,1,1,"linear_20.b_0"],"DA":[],"O":{"%":143,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[768],"NCHW",[],0]}},"OA":[1,0,1],"QA":[]},{"#":"p","A":[0,1,1,"linear_20.w_0"],"DA":[],"O":{"%":144,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[768,768],"NCHW",[],0]}},"OA":[1,0,1],"QA":[]},{"#":"p","A":[0,1,1,"linear_19.b_0"],"DA":[],"O":{"%":145,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[768],"NCHW",[],0]}},"OA":[1,0,1],"QA":[]},{"#":"p","A":[0,1,1,"linear_19.w_0"],"DA":[],"O":{"%":146,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[768,768],"NCHW",[],0]}},"OA":[1,0,1],"QA":[]},{"#":"p","A":[0,1,1,"linear_18.b_0"],"DA":[],"O":{"%":147,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[768],"NCHW",[],0]}},"OA":[1,0,1],"QA":[]},{"#":"p","A":[0,1,1,"linear_18.w_0"],"DA":[],"O":{"%":148,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[768,768],"NCHW",[],0]}},"OA":[1,0,1],"QA":[]},{"#":"p","A":[0,1,1,"layer_norm_6.b_0"],"DA":[],"O":{"%":149,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[768],"NCHW",[],0]}},"OA":[1,0,1],"QA":[]},{"#":"p","A":[0,1,1,"layer_norm_6.w_0"],"DA":[],"O":{"%":150,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[768],"NCHW",[],0]}},"OA":[1,0,1],"QA":[]},{"#":"p","A":[0,1,1,"layer_norm_5.b_0"],"DA":[],"O":{"%":151,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[768],"NCHW",[],0]}},"OA":[1,0,1],"QA":[]},{"#":"p","A":[0,1,1,"layer_norm_5.w_0"],"DA":[],"O":{"%":152,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[768],"NCHW",[],0]}},"OA":[1,0,1],"QA":[]},{"#":"p","A":[0,1,1,"linear_17.b_0"],"DA":[],"O":{"%":153,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[768],"NCHW",[],0]}},"OA":[1,0,1],"QA":[]},{"#":"p","A":[0,1,1,"linear_17.w_0"],"DA":[],"O":{"%":154,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[3072,768],"NCHW",[],0]}},"OA":[1,0,1],"QA":[]},{"#":"p","A":[0,1,1,"linear_16.b_0"],"DA":[],"O":{"%":155,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[3072],"NCHW",[],0]}},"OA":[1,0,1],"QA":[]},{"#":"p","A":[0,1,1,"linear_16.w_0"],"DA":[],"O":{"%":156,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[768,3072],"NCHW",[],0]}},"OA":[1,0,1],"QA":[]},{"#":"p","A":[0,1,1,"linear_15.b_0"],"DA":[],"O":{"%":157,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[768],"NCHW",[],0]}},"OA":[1,0,1],"QA":[]},{"#":"p","A":[0,1,1,"linear_15.w_0"],"DA":[],"O":{"%":158,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[768,768],"NCHW",[],0]}},"OA":[1,0,1],"QA":[]},{"#":"p","A":[0,1,1,"linear_14.b_0"],"DA":[],"O":{"%":159,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[768],"NCHW",[],0]}},"OA":[1,0,1],"QA":[]},{"#":"p","A":[0,1,1,"linear_14.w_0"],"DA":[],"O":{"%":160,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[768,768],"NCHW",[],0]}},"OA":[1,0,1],"QA":[]},{"#":"p","A":[0,1,1,"linear_13.b_0"],"DA":[],"O":{"%":161,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[768],"NCHW",[],0]}},"OA":[1,0,1],"QA":[]},{"#":"p","A":[0,1,1,"linear_13.w_0"],"DA":[],"O":{"%":162,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[768,768],"NCHW",[],0]}},"OA":[1,0,1],"QA":[]},{"#":"p","A":[0,1,1,"linear_12.b_0"],"DA":[],"O":{"%":163,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[768],"NCHW",[],0]}},"OA":[1,0,1],"QA":[]},{"#":"p","A":[0,1,1,"linear_12.w_0"],"DA":[],"O":{"%":164,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[768,768],"NCHW",[],0]}},"OA":[1,0,1],"QA":[]},{"#":"p","A":[0,1,1,"layer_norm_4.b_0"],"DA":[],"O":{"%":165,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[768],"NCHW",[],0]}},"OA":[1,0,1],"QA":[]},{"#":"p","A":[0,1,1,"layer_norm_4.w_0"],"DA":[],"O":{"%":166,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[768],"NCHW",[],0]}},"OA":[1,0,1],"QA":[]},{"#":"p","A":[0,1,1,"layer_norm_3.b_0"],"DA":[],"O":{"%":167,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[768],"NCHW",[],0]}},"OA":[1,0,1],"QA":[]},{"#":"p","A":[0,1,1,"layer_norm_3.w_0"],"DA":[],"O":{"%":168,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[768],"NCHW",[],0]}},"OA":[1,0,1],"QA":[]},{"#":"p","A":[0,1,1,"linear_11.b_0"],"DA":[],"O":{"%":169,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[768],"NCHW",[],0]}},"OA":[1,0,1],"QA":[]},{"#":"p","A":[0,1,1,"linear_11.w_0"],"DA":[],"O":{"%":170,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[3072,768],"NCHW",[],0]}},"OA":[1,0,1],"QA":[]},{"#":"p","A":[0,1,1,"linear_10.b_0"],"DA":[],"O":{"%":171,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[3072],"NCHW",[],0]}},"OA":[1,0,1],"QA":[]},{"#":"p","A":[0,1,1,"linear_10.w_0"],"DA":[],"O":{"%":172,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[768,3072],"NCHW",[],0]}},"OA":[1,0,1],"QA":[]},{"#":"p","A":[0,1,1,"linear_9.b_0"],"DA":[],"O":{"%":173,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[768],"NCHW",[],0]}},"OA":[1,0,1],"QA":[]},{"#":"p","A":[0,1,1,"linear_9.w_0"],"DA":[],"O":{"%":174,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[768,768],"NCHW",[],0]}},"OA":[1,0,1],"QA":[]},{"#":"p","A":[0,1,1,"linear_8.b_0"],"DA":[],"O":{"%":175,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[768],"NCHW",[],0]}},"OA":[1,0,1],"QA":[]},{"#":"p","A":[0,1,1,"linear_8.w_0"],"DA":[],"O":{"%":176,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[768,768],"NCHW",[],0]}},"OA":[1,0,1],"QA":[]},{"#":"p","A":[0,1,1,"linear_7.b_0"],"DA":[],"O":{"%":177,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[768],"NCHW",[],0]}},"OA":[1,0,1],"QA":[]},{"#":"p","A":[0,1,1,"linear_7.w_0"],"DA":[],"O":{"%":178,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[768,768],"NCHW",[],0]}},"OA":[1,0,1],"QA":[]},{"#":"p","A":[0,1,1,"linear_6.b_0"],"DA":[],"O":{"%":179,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[768],"NCHW",[],0]}},"OA":[1,0,1],"QA":[]},{"#":"p","A":[0,1,1,"linear_6.w_0"],"DA":[],"O":{"%":180,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[768,768],"NCHW",[],0]}},"OA":[1,0,1],"QA":[]},{"#":"p","A":[0,1,1,"layer_norm_2.b_0"],"DA":[],"O":{"%":181,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[768],"NCHW",[],0]}},"OA":[1,0,1],"QA":[]},{"#":"p","A":[0,1,1,"layer_norm_2.w_0"],"DA":[],"O":{"%":182,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[768],"NCHW",[],0]}},"OA":[1,0,1],"QA":[]},{"#":"p","A":[0,1,1,"layer_norm_1.b_0"],"DA":[],"O":{"%":183,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[768],"NCHW",[],0]}},"OA":[1,0,1],"QA":[]},{"#":"p","A":[0,1,1,"layer_norm_1.w_0"],"DA":[],"O":{"%":184,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[768],"NCHW",[],0]}},"OA":[1,0,1],"QA":[]},{"#":"p","A":[0,1,1,"linear_5.b_0"],"DA":[],"O":{"%":185,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[768],"NCHW",[],0]}},"OA":[1,0,1],"QA":[]},{"#":"p","A":[0,1,1,"linear_5.w_0"],"DA":[],"O":{"%":186,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[3072,768],"NCHW",[],0]}},"OA":[1,0,1],"QA":[]},{"#":"p","A":[0,1,1,"linear_4.b_0"],"DA":[],"O":{"%":187,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[3072],"NCHW",[],0]}},"OA":[1,0,1],"QA":[]},{"#":"p","A":[0,1,1,"linear_4.w_0"],"DA":[],"O":{"%":188,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[768,3072],"NCHW",[],0]}},"OA":[1,0,1],"QA":[]},{"#":"p","A":[0,1,1,"linear_3.b_0"],"DA":[],"O":{"%":189,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[768],"NCHW",[],0]}},"OA":[1,0,1],"QA":[]},{"#":"p","A":[0,1,1,"linear_3.w_0"],"DA":[],"O":{"%":190,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[768,768],"NCHW",[],0]}},"OA":[1,0,1],"QA":[]},{"#":"p","A":[0,1,1,"linear_2.b_0"],"DA":[],"O":{"%":191,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[768],"NCHW",[],0]}},"OA":[1,0,1],"QA":[]},{"#":"p","A":[0,1,1,"linear_2.w_0"],"DA":[],"O":{"%":192,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[768,768],"NCHW",[],0]}},"OA":[1,0,1],"QA":[]},{"#":"p","A":[0,1,1,"linear_1.b_0"],"DA":[],"O":{"%":193,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[768],"NCHW",[],0]}},"OA":[1,0,1],"QA":[]},{"#":"p","A":[0,1,1,"linear_1.w_0"],"DA":[],"O":{"%":194,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[768,768],"NCHW",[],0]}},"OA":[1,0,1],"QA":[]},{"#":"p","A":[0,1,1,"linear_0.b_0"],"DA":[],"O":{"%":195,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[768],"NCHW",[],0]}},"OA":[1,0,1],"QA":[]},{"#":"p","A":[0,1,1,"linear_0.w_0"],"DA":[],"O":{"%":196,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[768,768],"NCHW",[],0]}},"OA":[1,0,1],"QA":[]},{"#":"p","A":[0,1,1,"layer_norm_0.b_0"],"DA":[],"O":{"%":197,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[768],"NCHW",[],0]}},"OA":[1,0,1],"QA":[]},{"#":"p","A":[0,1,1,"layer_norm_0.w_0"],"DA":[],"O":{"%":198,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[768],"NCHW",[],0]}},"OA":[1,0,1],"QA":[]},{"#":"p","A":[0,1,1,"embedding_3.w_0"],"DA":[],"O":{"%":199,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[3,768],"NCHW",[],0]}},"OA":[1,0,1],"QA":[]},{"#":"p","A":[0,1,1,"embedding_2.w_0"],"DA":[],"O":{"%":200,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[4,768],"NCHW",[],0]}},"OA":[1,0,1],"QA":[]},{"#":"p","A":[0,1,1,"embedding_1.w_0"],"DA":[],"O":{"%":201,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[2048,768],"NCHW",[],0]}},"OA":[1,0,1],"QA":[]},{"#":"p","A":[0,1,1,"embedding_0.w_0"],"DA":[],"O":{"%":202,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[40000,768],"NCHW",[],0]}},"OA":[1,0,1],"QA":[]},{"#":"1.data","A":[{"AT":{"#":"0.a_str","D":"input_ids"},"N":"name"},{"AT":{"#":"1.a_intarray","D":[-1,512]},"N":"shape"},{"AT":{"#":"1.a_dtype","D":"int64"},"N":"dtype"},{"AT":{"#":"1.a_place","D":[0,0,""]},"N":"place"}],"I":[],"O":[{"%":203,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_i64"},[-1,512],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.data","A":[{"AT":{"#":"0.a_str","D":"token_type_ids"},"N":"name"},{"AT":{"#":"1.a_intarray","D":[-1,512]},"N":"shape"},{"AT":{"#":"1.a_dtype","D":"int64"},"N":"dtype"},{"AT":{"#":"1.a_place","D":[0,0,""]},"N":"place"}],"I":[],"O":[{"%":204,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_i64"},[-1,512],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.data","A":[{"AT":{"#":"0.a_str","D":"position_ids"},"N":"name"},{"AT":{"#":"1.a_intarray","D":[-1,512]},"N":"shape"},{"AT":{"#":"1.a_dtype","D":"int64"},"N":"dtype"},{"AT":{"#":"1.a_place","D":[0,0,""]},"N":"place"}],"I":[],"O":[{"%":205,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_i64"},[-1,512],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.data","A":[{"AT":{"#":"0.a_str","D":"attention_mask"},"N":"name"},{"AT":{"#":"1.a_intarray","D":[-1,512]},"N":"shape"},{"AT":{"#":"1.a_dtype","D":"float32"},"N":"dtype"},{"AT":{"#":"1.a_place","D":[0,0,""]},"N":"place"}],"I":[],"O":[{"%":206,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.full_int_array","A":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_i64","D":1},{"#":"0.a_i64","D":2}]},"N":"value"},{"AT":{"#":"1.a_dtype","D":"int64"},"N":"dtype"},{"AT":{"#":"1.a_place","D":[1,0,""]},"N":"place"},{"AT":{"#":"0.a_str","D":"/ErnieModel/"},"N":"struct_name"}],"I":[],"O":[{"%":207,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_i64"},[2],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":true}]},"N":"stop_gradient"}]},{"#":"1.unsqueeze","A":[{"AT":{"#":"0.a_str","D":"/ErnieModel/"},"N":"struct_name"}],"I":[{"%":206},{"%":207}],"O":[{"%":208,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,1,1,512],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.full","A":[{"AT":{"#":"1.a_intarray","D":[1]},"N":"shape"},{"AT":{"#":"0.a_f64","D":-1.0},"N":"value"},{"AT":{"#":"1.a_dtype","D":"float32"},"N":"dtype"},{"AT":{"#":"1.a_place","D":[1,0,""]},"N":"place"},{"AT":{"#":"0.a_str","D":"/ErnieModel/"},"N":"struct_name"}],"I":[],"O":[{"%":209,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[1],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":true}]},"N":"stop_gradient"}]},{"#":"1.scale","A":[{"AT":{"#":"0.a_f32","D":1.0},"N":"bias"},{"AT":{"#":"0.a_bool","D":true},"N":"bias_after_scale"},{"AT":{"#":"0.a_str","D":"/ErnieModel/"},"N":"struct_name"}],"I":[{"%":208},{"%":209}],"O":[{"%":210,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,1,1,512],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.full","A":[{"AT":{"#":"1.a_intarray","D":[1]},"N":"shape"},{"AT":{"#":"0.a_f64","D":-10000.0},"N":"value"},{"AT":{"#":"1.a_dtype","D":"float32"},"N":"dtype"},{"AT":{"#":"1.a_place","D":[1,0,""]},"N":"place"},{"AT":{"#":"0.a_str","D":"/ErnieModel/"},"N":"struct_name"}],"I":[],"O":[{"%":211,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[1],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":true}]},"N":"stop_gradient"}]},{"#":"1.scale","A":[{"AT":{"#":"0.a_f32","D":0.0},"N":"bias"},{"AT":{"#":"0.a_bool","D":true},"N":"bias_after_scale"},{"AT":{"#":"0.a_str","D":"/ErnieModel/"},"N":"struct_name"}],"I":[{"%":210},{"%":211}],"O":[{"%":212,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,1,1,512],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":true}]},"N":"stop_gradient"}]},{"#":"1.embedding","A":[{"AT":{"#":"0.a_i64","D":0},"N":"padding_idx"},{"AT":{"#":"0.a_bool","D":false},"N":"sparse"},{"AT":{"#":"0.a_str","D":"/ErnieModel/ErnieEmbeddings/Embedding/"},"N":"struct_name"}],"I":[{"%":203},{"%":202}],"O":[{"%":213,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512,768],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.shape64","A":[{"AT":{"#":"0.a_str","D":"/ErnieModel/ErnieEmbeddings/"},"N":"struct_name"}],"I":[{"%":213}],"O":[{"%":214,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_i64"},[3],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":true}]},"N":"stop_gradient"}]},{"#":"1.full_int_array","A":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_i64","D":0}]},"N":"value"},{"AT":{"#":"1.a_dtype","D":"int64"},"N":"dtype"},{"AT":{"#":"1.a_place","D":[1,0,""]},"N":"place"},{"AT":{"#":"0.a_str","D":"/ErnieModel/ErnieEmbeddings/"},"N":"struct_name"}],"I":[],"O":[{"%":215,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_i64"},[1],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":true}]},"N":"stop_gradient"}]},{"#":"1.full_int_array","A":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_i64","D":1}]},"N":"value"},{"AT":{"#":"1.a_dtype","D":"int64"},"N":"dtype"},{"AT":{"#":"1.a_place","D":[1,0,""]},"N":"place"},{"AT":{"#":"0.a_str","D":"/ErnieModel/ErnieEmbeddings/"},"N":"struct_name"}],"I":[],"O":[{"%":216,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_i64"},[1],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":true}]},"N":"stop_gradient"}]},{"#":"1.slice","A":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_i64","D":0}]},"N":"axes"},{"AT":{"#":"0.a_array","D":[{"#":"0.a_i64","D":1}]},"N":"infer_flags"},{"AT":{"#":"0.a_array","D":[{"#":"0.a_i64","D":0}]},"N":"decrease_axis"},{"AT":{"#":"0.a_str","D":"/ErnieModel/ErnieEmbeddings/"},"N":"struct_name"}],"I":[{"%":214},{"%":215},{"%":216}],"O":[{"%":217,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_i64"},[],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":true}]},"N":"stop_gradient"}]},{"#":"1.embedding","A":[{"AT":{"#":"0.a_i64","D":-1},"N":"padding_idx"},{"AT":{"#":"0.a_bool","D":false},"N":"sparse"},{"AT":{"#":"0.a_str","D":"/ErnieModel/ErnieEmbeddings/Embedding_1/"},"N":"struct_name"}],"I":[{"%":205},{"%":201}],"O":[{"%":218,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512,768],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.add","A":[{"AT":{"#":"0.a_str","D":"/ErnieModel/ErnieEmbeddings/"},"N":"struct_name"}],"I":[{"%":213},{"%":218}],"O":[{"%":219,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512,768],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.embedding","A":[{"AT":{"#":"0.a_i64","D":-1},"N":"padding_idx"},{"AT":{"#":"0.a_bool","D":false},"N":"sparse"},{"AT":{"#":"0.a_str","D":"/ErnieModel/ErnieEmbeddings/Embedding_2/"},"N":"struct_name"}],"I":[{"%":204},{"%":200}],"O":[{"%":220,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512,768],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.add","A":[{"AT":{"#":"0.a_str","D":"/ErnieModel/ErnieEmbeddings/"},"N":"struct_name"}],"I":[{"%":219},{"%":220}],"O":[{"%":221,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512,768],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.full","A":[{"AT":{"#":"1.a_intarray","D":[]},"N":"shape"},{"AT":{"#":"0.a_f64","D":512.0},"N":"value"},{"AT":{"#":"1.a_dtype","D":"int64"},"N":"dtype"},{"AT":{"#":"1.a_place","D":[1,0,""]},"N":"place"},{"AT":{"#":"0.a_str","D":"/ErnieModel/ErnieEmbeddings/"},"N":"struct_name"}],"I":[],"O":[{"%":222,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_i64"},[],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":true}]},"N":"stop_gradient"}]},{"#":"0.combine","A":[{"AT":{"#":"0.a_str","D":"/ErnieModel/ErnieEmbeddings/"},"N":"struct_name"}],"I":[{"%":217},{"%":222}],"O":[{"%":223,"TT":{"#":"0.t_vec","D":[{"#":"0.t_dtensor","D":[{"#":"0.t_i64"},[],"NCHW",[],0]},{"#":"0.t_dtensor","D":[{"#":"0.t_i64"},[],"NCHW",[],0]}]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":true}]},"N":"stop_gradient"}]},{"#":"1.stack","A":[{"AT":{"#":"0.a_i32","D":0},"N":"axis"},{"AT":{"#":"0.a_str","D":"/ErnieModel/ErnieEmbeddings/"},"N":"struct_name"}],"I":[{"%":223}],"O":[{"%":224,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_i64"},[2],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":true}]},"N":"stop_gradient"}]},{"#":"1.full","A":[{"AT":{"#":"1.a_intarray","D":[1]},"N":"shape"},{"AT":{"#":"0.a_f64","D":1.0},"N":"value"},{"AT":{"#":"1.a_dtype","D":"float32"},"N":"dtype"},{"AT":{"#":"1.a_place","D":[1,0,""]},"N":"place"},{"AT":{"#":"0.a_str","D":"/ErnieModel/ErnieEmbeddings/"},"N":"struct_name"}],"I":[],"O":[{"%":225,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[1],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":true}]},"N":"stop_gradient"}]},{"#":"1.full_with_tensor","A":[{"AT":{"#":"1.a_dtype","D":"int64"},"N":"dtype"},{"AT":{"#":"0.a_str","D":"/ErnieModel/ErnieEmbeddings/"},"N":"struct_name"}],"I":[{"%":225},{"%":224}],"O":[{"%":226,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_i64"},[-1,512],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":true}]},"N":"stop_gradient"}]},{"#":"1.full","A":[{"AT":{"#":"1.a_intarray","D":[1]},"N":"shape"},{"AT":{"#":"0.a_f64","D":0.0},"N":"value"},{"AT":{"#":"1.a_dtype","D":"float32"},"N":"dtype"},{"AT":{"#":"1.a_place","D":[1,0,""]},"N":"place"},{"AT":{"#":"0.a_str","D":"/ErnieModel/ErnieEmbeddings/"},"N":"struct_name"}],"I":[],"O":[{"%":227,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[1],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":true}]},"N":"stop_gradient"}]},{"#":"1.scale","A":[{"AT":{"#":"0.a_f32","D":0.0},"N":"bias"},{"AT":{"#":"0.a_bool","D":true},"N":"bias_after_scale"},{"AT":{"#":"0.a_str","D":"/ErnieModel/ErnieEmbeddings/"},"N":"struct_name"}],"I":[{"%":226},{"%":227}],"O":[{"%":228,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_i64"},[-1,512],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":true}]},"N":"stop_gradient"}]},{"#":"1.embedding","A":[{"AT":{"#":"0.a_i64","D":-1},"N":"padding_idx"},{"AT":{"#":"0.a_bool","D":false},"N":"sparse"},{"AT":{"#":"0.a_str","D":"/ErnieModel/ErnieEmbeddings/Embedding_3/"},"N":"struct_name"}],"I":[{"%":228},{"%":199}],"O":[{"%":229,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512,768],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.add","A":[{"AT":{"#":"0.a_str","D":"/ErnieModel/ErnieEmbeddings/"},"N":"struct_name"}],"I":[{"%":221},{"%":229}],"O":[{"%":230,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512,768],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.layer_norm","A":[{"AT":{"#":"0.a_f32","D":9.999999960041972e-13},"N":"epsilon"},{"AT":{"#":"0.a_i32","D":2},"N":"begin_norm_axis"},{"AT":{"#":"0.a_str","D":"/ErnieModel/ErnieEmbeddings/LayerNorm/"},"N":"struct_name"}],"I":[{"%":230},{"%":198},{"%":197}],"O":[{"%":231,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512,768],"NCHW",[],0]}},{"%":232,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512],"NCHW",[],0]}},{"%":233,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false},{"#":"0.a_bool","D":false},{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.full","A":[{"AT":{"#":"1.a_intarray","D":[1]},"N":"shape"},{"AT":{"#":"0.a_f64","D":0.10000000149011612},"N":"value"},{"AT":{"#":"1.a_dtype","D":"float32"},"N":"dtype"},{"AT":{"#":"1.a_place","D":[1,0,""]},"N":"place"},{"AT":{"#":"0.a_str","D":"/ErnieModel/ErnieEmbeddings/Dropout/"},"N":"struct_name"}],"I":[],"O":[{"%":234,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[1],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":true}]},"N":"stop_gradient"}]},{"#":"1.dropout","A":[{"AT":{"#":"0.a_bool","D":true},"N":"is_test"},{"AT":{"#":"0.a_str","D":"upscale_in_train"},"N":"mode"},{"AT":{"#":"0.a_i32","D":0},"N":"seed"},{"AT":{"#":"0.a_bool","D":false},"N":"fix_seed"},{"AT":{"#":"0.a_str","D":"/ErnieModel/ErnieEmbeddings/Dropout/"},"N":"struct_name"}],"I":[{"%":231},{"%":0},{"%":234}],"O":[{"%":235,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512,768],"NCHW",[],0]}},{"%":236,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_ui8"},[-1,512,768],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false},{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.matmul","A":[{"AT":{"#":"0.a_bool","D":false},"N":"transpose_x"},{"AT":{"#":"0.a_bool","D":false},"N":"transpose_y"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer/MultiHeadAttention/Linear/"},"N":"struct_name"}],"I":[{"%":235},{"%":196}],"O":[{"%":237,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512,768],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.add","A":[{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer/MultiHeadAttention/Linear/"},"N":"struct_name"}],"I":[{"%":237},{"%":195}],"O":[{"%":238,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512,768],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.full_int_array","A":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_i64","D":0},{"#":"0.a_i64","D":0},{"#":"0.a_i64","D":12},{"#":"0.a_i64","D":64}]},"N":"value"},{"AT":{"#":"1.a_dtype","D":"int64"},"N":"dtype"},{"AT":{"#":"1.a_place","D":[1,0,""]},"N":"place"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer/MultiHeadAttention/"},"N":"struct_name"}],"I":[],"O":[{"%":239,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_i64"},[4],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":true}]},"N":"stop_gradient"}]},{"#":"1.reshape","A":[{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer/MultiHeadAttention/"},"N":"struct_name"}],"I":[{"%":238},{"%":239}],"O":[{"%":240,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512,12,64],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.transpose","A":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_i32","D":0},{"#":"0.a_i32","D":2},{"#":"0.a_i32","D":1},{"#":"0.a_i32","D":3}]},"N":"perm"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer/MultiHeadAttention/"},"N":"struct_name"}],"I":[{"%":240}],"O":[{"%":241,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,12,512,64],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.matmul","A":[{"AT":{"#":"0.a_bool","D":false},"N":"transpose_x"},{"AT":{"#":"0.a_bool","D":false},"N":"transpose_y"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer/MultiHeadAttention/Linear_1/"},"N":"struct_name"}],"I":[{"%":235},{"%":194}],"O":[{"%":242,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512,768],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.add","A":[{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer/MultiHeadAttention/Linear_1/"},"N":"struct_name"}],"I":[{"%":242},{"%":193}],"O":[{"%":243,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512,768],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.matmul","A":[{"AT":{"#":"0.a_bool","D":false},"N":"transpose_x"},{"AT":{"#":"0.a_bool","D":false},"N":"transpose_y"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer/MultiHeadAttention/Linear_2/"},"N":"struct_name"}],"I":[{"%":235},{"%":192}],"O":[{"%":244,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512,768],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.add","A":[{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer/MultiHeadAttention/Linear_2/"},"N":"struct_name"}],"I":[{"%":244},{"%":191}],"O":[{"%":245,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512,768],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.full_int_array","A":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_i64","D":0},{"#":"0.a_i64","D":0},{"#":"0.a_i64","D":12},{"#":"0.a_i64","D":64}]},"N":"value"},{"AT":{"#":"1.a_dtype","D":"int64"},"N":"dtype"},{"AT":{"#":"1.a_place","D":[1,0,""]},"N":"place"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer/MultiHeadAttention/"},"N":"struct_name"}],"I":[],"O":[{"%":246,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_i64"},[4],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":true}]},"N":"stop_gradient"}]},{"#":"1.reshape","A":[{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer/MultiHeadAttention/"},"N":"struct_name"}],"I":[{"%":243},{"%":246}],"O":[{"%":247,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512,12,64],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.transpose","A":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_i32","D":0},{"#":"0.a_i32","D":2},{"#":"0.a_i32","D":1},{"#":"0.a_i32","D":3}]},"N":"perm"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer/MultiHeadAttention/"},"N":"struct_name"}],"I":[{"%":247}],"O":[{"%":248,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,12,512,64],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.full_int_array","A":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_i64","D":0},{"#":"0.a_i64","D":0},{"#":"0.a_i64","D":12},{"#":"0.a_i64","D":64}]},"N":"value"},{"AT":{"#":"1.a_dtype","D":"int64"},"N":"dtype"},{"AT":{"#":"1.a_place","D":[1,0,""]},"N":"place"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer/MultiHeadAttention/"},"N":"struct_name"}],"I":[],"O":[{"%":249,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_i64"},[4],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":true}]},"N":"stop_gradient"}]},{"#":"1.reshape","A":[{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer/MultiHeadAttention/"},"N":"struct_name"}],"I":[{"%":245},{"%":249}],"O":[{"%":250,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512,12,64],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.transpose","A":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_i32","D":0},{"#":"0.a_i32","D":2},{"#":"0.a_i32","D":1},{"#":"0.a_i32","D":3}]},"N":"perm"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer/MultiHeadAttention/"},"N":"struct_name"}],"I":[{"%":250}],"O":[{"%":251,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,12,512,64],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.full","A":[{"AT":{"#":"1.a_intarray","D":[1]},"N":"shape"},{"AT":{"#":"0.a_f64","D":0.125},"N":"value"},{"AT":{"#":"1.a_dtype","D":"float32"},"N":"dtype"},{"AT":{"#":"1.a_place","D":[1,0,""]},"N":"place"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer/MultiHeadAttention/"},"N":"struct_name"}],"I":[],"O":[{"%":252,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[1],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":true}]},"N":"stop_gradient"}]},{"#":"1.scale","A":[{"AT":{"#":"0.a_f32","D":0.0},"N":"bias"},{"AT":{"#":"0.a_bool","D":true},"N":"bias_after_scale"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer/MultiHeadAttention/"},"N":"struct_name"}],"I":[{"%":241},{"%":252}],"O":[{"%":253,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,12,512,64],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.matmul","A":[{"AT":{"#":"0.a_bool","D":false},"N":"transpose_x"},{"AT":{"#":"0.a_bool","D":true},"N":"transpose_y"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer/MultiHeadAttention/"},"N":"struct_name"}],"I":[{"%":253},{"%":248}],"O":[{"%":254,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,12,512,512],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.add","A":[{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer/MultiHeadAttention/"},"N":"struct_name"}],"I":[{"%":254},{"%":212}],"O":[{"%":255,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,12,512,512],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.softmax","A":[{"AT":{"#":"0.a_i32","D":-1},"N":"axis"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer/MultiHeadAttention/"},"N":"struct_name"}],"I":[{"%":255}],"O":[{"%":256,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,12,512,512],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.full","A":[{"AT":{"#":"1.a_intarray","D":[1]},"N":"shape"},{"AT":{"#":"0.a_f64","D":0.10000000149011612},"N":"value"},{"AT":{"#":"1.a_dtype","D":"float32"},"N":"dtype"},{"AT":{"#":"1.a_place","D":[1,0,""]},"N":"place"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer/MultiHeadAttention/"},"N":"struct_name"}],"I":[],"O":[{"%":257,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[1],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":true}]},"N":"stop_gradient"}]},{"#":"1.dropout","A":[{"AT":{"#":"0.a_bool","D":true},"N":"is_test"},{"AT":{"#":"0.a_str","D":"upscale_in_train"},"N":"mode"},{"AT":{"#":"0.a_i32","D":0},"N":"seed"},{"AT":{"#":"0.a_bool","D":false},"N":"fix_seed"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer/MultiHeadAttention/"},"N":"struct_name"}],"I":[{"%":256},{"%":0},{"%":257}],"O":[{"%":258,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,12,512,512],"NCHW",[],0]}},{"%":259,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_ui8"},[-1,12,512,512],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false},{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.matmul","A":[{"AT":{"#":"0.a_bool","D":false},"N":"transpose_x"},{"AT":{"#":"0.a_bool","D":false},"N":"transpose_y"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer/MultiHeadAttention/"},"N":"struct_name"}],"I":[{"%":258},{"%":251}],"O":[{"%":260,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,12,512,64],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.transpose","A":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_i32","D":0},{"#":"0.a_i32","D":2},{"#":"0.a_i32","D":1},{"#":"0.a_i32","D":3}]},"N":"perm"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer/MultiHeadAttention/"},"N":"struct_name"}],"I":[{"%":260}],"O":[{"%":261,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512,12,64],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.full_int_array","A":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_i64","D":0},{"#":"0.a_i64","D":0},{"#":"0.a_i64","D":768}]},"N":"value"},{"AT":{"#":"1.a_dtype","D":"int64"},"N":"dtype"},{"AT":{"#":"1.a_place","D":[1,0,""]},"N":"place"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer/MultiHeadAttention/"},"N":"struct_name"}],"I":[],"O":[{"%":262,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_i64"},[3],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":true}]},"N":"stop_gradient"}]},{"#":"1.reshape","A":[{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer/MultiHeadAttention/"},"N":"struct_name"}],"I":[{"%":261},{"%":262}],"O":[{"%":263,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512,768],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.matmul","A":[{"AT":{"#":"0.a_bool","D":false},"N":"transpose_x"},{"AT":{"#":"0.a_bool","D":false},"N":"transpose_y"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer/MultiHeadAttention/Linear_3/"},"N":"struct_name"}],"I":[{"%":263},{"%":190}],"O":[{"%":264,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512,768],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.add","A":[{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer/MultiHeadAttention/Linear_3/"},"N":"struct_name"}],"I":[{"%":264},{"%":189}],"O":[{"%":265,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512,768],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.full","A":[{"AT":{"#":"1.a_intarray","D":[1]},"N":"shape"},{"AT":{"#":"0.a_f64","D":0.10000000149011612},"N":"value"},{"AT":{"#":"1.a_dtype","D":"float32"},"N":"dtype"},{"AT":{"#":"1.a_place","D":[1,0,""]},"N":"place"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer/Dropout/"},"N":"struct_name"}],"I":[],"O":[{"%":266,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[1],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":true}]},"N":"stop_gradient"}]},{"#":"1.dropout","A":[{"AT":{"#":"0.a_bool","D":true},"N":"is_test"},{"AT":{"#":"0.a_str","D":"upscale_in_train"},"N":"mode"},{"AT":{"#":"0.a_i32","D":0},"N":"seed"},{"AT":{"#":"0.a_bool","D":false},"N":"fix_seed"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer/Dropout/"},"N":"struct_name"}],"I":[{"%":265},{"%":0},{"%":266}],"O":[{"%":267,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512,768],"NCHW",[],0]}},{"%":268,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_ui8"},[-1,512,768],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false},{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.add","A":[{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer/"},"N":"struct_name"}],"I":[{"%":235},{"%":267}],"O":[{"%":269,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512,768],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.layer_norm","A":[{"AT":{"#":"0.a_f32","D":9.999999960041972e-13},"N":"epsilon"},{"AT":{"#":"0.a_i32","D":2},"N":"begin_norm_axis"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer/LayerNorm/"},"N":"struct_name"}],"I":[{"%":269},{"%":184},{"%":183}],"O":[{"%":270,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512,768],"NCHW",[],0]}},{"%":271,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512],"NCHW",[],0]}},{"%":272,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false},{"#":"0.a_bool","D":false},{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.matmul","A":[{"AT":{"#":"0.a_bool","D":false},"N":"transpose_x"},{"AT":{"#":"0.a_bool","D":false},"N":"transpose_y"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer/Linear/"},"N":"struct_name"}],"I":[{"%":270},{"%":188}],"O":[{"%":273,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512,3072],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.add","A":[{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer/Linear/"},"N":"struct_name"}],"I":[{"%":273},{"%":187}],"O":[{"%":274,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512,3072],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.gelu","A":[{"AT":{"#":"0.a_bool","D":false},"N":"approximate"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer/"},"N":"struct_name"}],"I":[{"%":274}],"O":[{"%":275,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512,3072],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.matmul","A":[{"AT":{"#":"0.a_bool","D":false},"N":"transpose_x"},{"AT":{"#":"0.a_bool","D":false},"N":"transpose_y"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer/Linear_1/"},"N":"struct_name"}],"I":[{"%":275},{"%":186}],"O":[{"%":276,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512,768],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.add","A":[{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer/Linear_1/"},"N":"struct_name"}],"I":[{"%":276},{"%":185}],"O":[{"%":277,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512,768],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.full","A":[{"AT":{"#":"1.a_intarray","D":[1]},"N":"shape"},{"AT":{"#":"0.a_f64","D":0.10000000149011612},"N":"value"},{"AT":{"#":"1.a_dtype","D":"float32"},"N":"dtype"},{"AT":{"#":"1.a_place","D":[1,0,""]},"N":"place"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer/Dropout_2/"},"N":"struct_name"}],"I":[],"O":[{"%":278,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[1],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":true}]},"N":"stop_gradient"}]},{"#":"1.dropout","A":[{"AT":{"#":"0.a_bool","D":true},"N":"is_test"},{"AT":{"#":"0.a_str","D":"upscale_in_train"},"N":"mode"},{"AT":{"#":"0.a_i32","D":0},"N":"seed"},{"AT":{"#":"0.a_bool","D":false},"N":"fix_seed"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer/Dropout_2/"},"N":"struct_name"}],"I":[{"%":277},{"%":0},{"%":278}],"O":[{"%":279,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512,768],"NCHW",[],0]}},{"%":280,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_ui8"},[-1,512,768],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false},{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.add","A":[{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer/"},"N":"struct_name"}],"I":[{"%":270},{"%":279}],"O":[{"%":281,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512,768],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.layer_norm","A":[{"AT":{"#":"0.a_f32","D":9.999999960041972e-13},"N":"epsilon"},{"AT":{"#":"0.a_i32","D":2},"N":"begin_norm_axis"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer/LayerNorm_1/"},"N":"struct_name"}],"I":[{"%":281},{"%":182},{"%":181}],"O":[{"%":282,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512,768],"NCHW",[],0]}},{"%":283,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512],"NCHW",[],0]}},{"%":284,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false},{"#":"0.a_bool","D":false},{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.matmul","A":[{"AT":{"#":"0.a_bool","D":false},"N":"transpose_x"},{"AT":{"#":"0.a_bool","D":false},"N":"transpose_y"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_1/MultiHeadAttention/Linear/"},"N":"struct_name"}],"I":[{"%":282},{"%":180}],"O":[{"%":285,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512,768],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.add","A":[{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_1/MultiHeadAttention/Linear/"},"N":"struct_name"}],"I":[{"%":285},{"%":179}],"O":[{"%":286,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512,768],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.full_int_array","A":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_i64","D":0},{"#":"0.a_i64","D":0},{"#":"0.a_i64","D":12},{"#":"0.a_i64","D":64}]},"N":"value"},{"AT":{"#":"1.a_dtype","D":"int64"},"N":"dtype"},{"AT":{"#":"1.a_place","D":[1,0,""]},"N":"place"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_1/MultiHeadAttention/"},"N":"struct_name"}],"I":[],"O":[{"%":287,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_i64"},[4],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":true}]},"N":"stop_gradient"}]},{"#":"1.reshape","A":[{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_1/MultiHeadAttention/"},"N":"struct_name"}],"I":[{"%":286},{"%":287}],"O":[{"%":288,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512,12,64],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.transpose","A":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_i32","D":0},{"#":"0.a_i32","D":2},{"#":"0.a_i32","D":1},{"#":"0.a_i32","D":3}]},"N":"perm"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_1/MultiHeadAttention/"},"N":"struct_name"}],"I":[{"%":288}],"O":[{"%":289,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,12,512,64],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.matmul","A":[{"AT":{"#":"0.a_bool","D":false},"N":"transpose_x"},{"AT":{"#":"0.a_bool","D":false},"N":"transpose_y"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_1/MultiHeadAttention/Linear_1/"},"N":"struct_name"}],"I":[{"%":282},{"%":178}],"O":[{"%":290,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512,768],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.add","A":[{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_1/MultiHeadAttention/Linear_1/"},"N":"struct_name"}],"I":[{"%":290},{"%":177}],"O":[{"%":291,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512,768],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.matmul","A":[{"AT":{"#":"0.a_bool","D":false},"N":"transpose_x"},{"AT":{"#":"0.a_bool","D":false},"N":"transpose_y"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_1/MultiHeadAttention/Linear_2/"},"N":"struct_name"}],"I":[{"%":282},{"%":176}],"O":[{"%":292,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512,768],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.add","A":[{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_1/MultiHeadAttention/Linear_2/"},"N":"struct_name"}],"I":[{"%":292},{"%":175}],"O":[{"%":293,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512,768],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.full_int_array","A":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_i64","D":0},{"#":"0.a_i64","D":0},{"#":"0.a_i64","D":12},{"#":"0.a_i64","D":64}]},"N":"value"},{"AT":{"#":"1.a_dtype","D":"int64"},"N":"dtype"},{"AT":{"#":"1.a_place","D":[1,0,""]},"N":"place"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_1/MultiHeadAttention/"},"N":"struct_name"}],"I":[],"O":[{"%":294,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_i64"},[4],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":true}]},"N":"stop_gradient"}]},{"#":"1.reshape","A":[{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_1/MultiHeadAttention/"},"N":"struct_name"}],"I":[{"%":291},{"%":294}],"O":[{"%":295,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512,12,64],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.transpose","A":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_i32","D":0},{"#":"0.a_i32","D":2},{"#":"0.a_i32","D":1},{"#":"0.a_i32","D":3}]},"N":"perm"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_1/MultiHeadAttention/"},"N":"struct_name"}],"I":[{"%":295}],"O":[{"%":296,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,12,512,64],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.full_int_array","A":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_i64","D":0},{"#":"0.a_i64","D":0},{"#":"0.a_i64","D":12},{"#":"0.a_i64","D":64}]},"N":"value"},{"AT":{"#":"1.a_dtype","D":"int64"},"N":"dtype"},{"AT":{"#":"1.a_place","D":[1,0,""]},"N":"place"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_1/MultiHeadAttention/"},"N":"struct_name"}],"I":[],"O":[{"%":297,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_i64"},[4],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":true}]},"N":"stop_gradient"}]},{"#":"1.reshape","A":[{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_1/MultiHeadAttention/"},"N":"struct_name"}],"I":[{"%":293},{"%":297}],"O":[{"%":298,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512,12,64],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.transpose","A":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_i32","D":0},{"#":"0.a_i32","D":2},{"#":"0.a_i32","D":1},{"#":"0.a_i32","D":3}]},"N":"perm"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_1/MultiHeadAttention/"},"N":"struct_name"}],"I":[{"%":298}],"O":[{"%":299,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,12,512,64],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.full","A":[{"AT":{"#":"1.a_intarray","D":[1]},"N":"shape"},{"AT":{"#":"0.a_f64","D":0.125},"N":"value"},{"AT":{"#":"1.a_dtype","D":"float32"},"N":"dtype"},{"AT":{"#":"1.a_place","D":[1,0,""]},"N":"place"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_1/MultiHeadAttention/"},"N":"struct_name"}],"I":[],"O":[{"%":300,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[1],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":true}]},"N":"stop_gradient"}]},{"#":"1.scale","A":[{"AT":{"#":"0.a_f32","D":0.0},"N":"bias"},{"AT":{"#":"0.a_bool","D":true},"N":"bias_after_scale"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_1/MultiHeadAttention/"},"N":"struct_name"}],"I":[{"%":289},{"%":300}],"O":[{"%":301,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,12,512,64],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.matmul","A":[{"AT":{"#":"0.a_bool","D":false},"N":"transpose_x"},{"AT":{"#":"0.a_bool","D":true},"N":"transpose_y"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_1/MultiHeadAttention/"},"N":"struct_name"}],"I":[{"%":301},{"%":296}],"O":[{"%":302,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,12,512,512],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.add","A":[{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_1/MultiHeadAttention/"},"N":"struct_name"}],"I":[{"%":302},{"%":212}],"O":[{"%":303,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,12,512,512],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.softmax","A":[{"AT":{"#":"0.a_i32","D":-1},"N":"axis"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_1/MultiHeadAttention/"},"N":"struct_name"}],"I":[{"%":303}],"O":[{"%":304,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,12,512,512],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.full","A":[{"AT":{"#":"1.a_intarray","D":[1]},"N":"shape"},{"AT":{"#":"0.a_f64","D":0.10000000149011612},"N":"value"},{"AT":{"#":"1.a_dtype","D":"float32"},"N":"dtype"},{"AT":{"#":"1.a_place","D":[1,0,""]},"N":"place"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_1/MultiHeadAttention/"},"N":"struct_name"}],"I":[],"O":[{"%":305,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[1],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":true}]},"N":"stop_gradient"}]},{"#":"1.dropout","A":[{"AT":{"#":"0.a_bool","D":true},"N":"is_test"},{"AT":{"#":"0.a_str","D":"upscale_in_train"},"N":"mode"},{"AT":{"#":"0.a_i32","D":0},"N":"seed"},{"AT":{"#":"0.a_bool","D":false},"N":"fix_seed"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_1/MultiHeadAttention/"},"N":"struct_name"}],"I":[{"%":304},{"%":0},{"%":305}],"O":[{"%":306,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,12,512,512],"NCHW",[],0]}},{"%":307,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_ui8"},[-1,12,512,512],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false},{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.matmul","A":[{"AT":{"#":"0.a_bool","D":false},"N":"transpose_x"},{"AT":{"#":"0.a_bool","D":false},"N":"transpose_y"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_1/MultiHeadAttention/"},"N":"struct_name"}],"I":[{"%":306},{"%":299}],"O":[{"%":308,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,12,512,64],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.transpose","A":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_i32","D":0},{"#":"0.a_i32","D":2},{"#":"0.a_i32","D":1},{"#":"0.a_i32","D":3}]},"N":"perm"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_1/MultiHeadAttention/"},"N":"struct_name"}],"I":[{"%":308}],"O":[{"%":309,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512,12,64],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.full_int_array","A":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_i64","D":0},{"#":"0.a_i64","D":0},{"#":"0.a_i64","D":768}]},"N":"value"},{"AT":{"#":"1.a_dtype","D":"int64"},"N":"dtype"},{"AT":{"#":"1.a_place","D":[1,0,""]},"N":"place"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_1/MultiHeadAttention/"},"N":"struct_name"}],"I":[],"O":[{"%":310,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_i64"},[3],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":true}]},"N":"stop_gradient"}]},{"#":"1.reshape","A":[{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_1/MultiHeadAttention/"},"N":"struct_name"}],"I":[{"%":309},{"%":310}],"O":[{"%":311,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512,768],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.matmul","A":[{"AT":{"#":"0.a_bool","D":false},"N":"transpose_x"},{"AT":{"#":"0.a_bool","D":false},"N":"transpose_y"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_1/MultiHeadAttention/Linear_3/"},"N":"struct_name"}],"I":[{"%":311},{"%":174}],"O":[{"%":312,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512,768],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.add","A":[{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_1/MultiHeadAttention/Linear_3/"},"N":"struct_name"}],"I":[{"%":312},{"%":173}],"O":[{"%":313,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512,768],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.full","A":[{"AT":{"#":"1.a_intarray","D":[1]},"N":"shape"},{"AT":{"#":"0.a_f64","D":0.10000000149011612},"N":"value"},{"AT":{"#":"1.a_dtype","D":"float32"},"N":"dtype"},{"AT":{"#":"1.a_place","D":[1,0,""]},"N":"place"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_1/Dropout/"},"N":"struct_name"}],"I":[],"O":[{"%":314,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[1],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":true}]},"N":"stop_gradient"}]},{"#":"1.dropout","A":[{"AT":{"#":"0.a_bool","D":true},"N":"is_test"},{"AT":{"#":"0.a_str","D":"upscale_in_train"},"N":"mode"},{"AT":{"#":"0.a_i32","D":0},"N":"seed"},{"AT":{"#":"0.a_bool","D":false},"N":"fix_seed"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_1/Dropout/"},"N":"struct_name"}],"I":[{"%":313},{"%":0},{"%":314}],"O":[{"%":315,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512,768],"NCHW",[],0]}},{"%":316,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_ui8"},[-1,512,768],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false},{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.add","A":[{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_1/"},"N":"struct_name"}],"I":[{"%":282},{"%":315}],"O":[{"%":317,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512,768],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.layer_norm","A":[{"AT":{"#":"0.a_f32","D":9.999999960041972e-13},"N":"epsilon"},{"AT":{"#":"0.a_i32","D":2},"N":"begin_norm_axis"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_1/LayerNorm/"},"N":"struct_name"}],"I":[{"%":317},{"%":168},{"%":167}],"O":[{"%":318,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512,768],"NCHW",[],0]}},{"%":319,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512],"NCHW",[],0]}},{"%":320,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false},{"#":"0.a_bool","D":false},{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.matmul","A":[{"AT":{"#":"0.a_bool","D":false},"N":"transpose_x"},{"AT":{"#":"0.a_bool","D":false},"N":"transpose_y"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_1/Linear/"},"N":"struct_name"}],"I":[{"%":318},{"%":172}],"O":[{"%":321,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512,3072],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.add","A":[{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_1/Linear/"},"N":"struct_name"}],"I":[{"%":321},{"%":171}],"O":[{"%":322,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512,3072],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.gelu","A":[{"AT":{"#":"0.a_bool","D":false},"N":"approximate"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_1/"},"N":"struct_name"}],"I":[{"%":322}],"O":[{"%":323,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512,3072],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.matmul","A":[{"AT":{"#":"0.a_bool","D":false},"N":"transpose_x"},{"AT":{"#":"0.a_bool","D":false},"N":"transpose_y"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_1/Linear_1/"},"N":"struct_name"}],"I":[{"%":323},{"%":170}],"O":[{"%":324,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512,768],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.add","A":[{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_1/Linear_1/"},"N":"struct_name"}],"I":[{"%":324},{"%":169}],"O":[{"%":325,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512,768],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.full","A":[{"AT":{"#":"1.a_intarray","D":[1]},"N":"shape"},{"AT":{"#":"0.a_f64","D":0.10000000149011612},"N":"value"},{"AT":{"#":"1.a_dtype","D":"float32"},"N":"dtype"},{"AT":{"#":"1.a_place","D":[1,0,""]},"N":"place"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_1/Dropout_2/"},"N":"struct_name"}],"I":[],"O":[{"%":326,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[1],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":true}]},"N":"stop_gradient"}]},{"#":"1.dropout","A":[{"AT":{"#":"0.a_bool","D":true},"N":"is_test"},{"AT":{"#":"0.a_str","D":"upscale_in_train"},"N":"mode"},{"AT":{"#":"0.a_i32","D":0},"N":"seed"},{"AT":{"#":"0.a_bool","D":false},"N":"fix_seed"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_1/Dropout_2/"},"N":"struct_name"}],"I":[{"%":325},{"%":0},{"%":326}],"O":[{"%":327,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512,768],"NCHW",[],0]}},{"%":328,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_ui8"},[-1,512,768],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false},{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.add","A":[{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_1/"},"N":"struct_name"}],"I":[{"%":318},{"%":327}],"O":[{"%":329,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512,768],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.layer_norm","A":[{"AT":{"#":"0.a_f32","D":9.999999960041972e-13},"N":"epsilon"},{"AT":{"#":"0.a_i32","D":2},"N":"begin_norm_axis"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_1/LayerNorm_1/"},"N":"struct_name"}],"I":[{"%":329},{"%":166},{"%":165}],"O":[{"%":330,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512,768],"NCHW",[],0]}},{"%":331,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512],"NCHW",[],0]}},{"%":332,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false},{"#":"0.a_bool","D":false},{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.matmul","A":[{"AT":{"#":"0.a_bool","D":false},"N":"transpose_x"},{"AT":{"#":"0.a_bool","D":false},"N":"transpose_y"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_2/MultiHeadAttention/Linear/"},"N":"struct_name"}],"I":[{"%":330},{"%":164}],"O":[{"%":333,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512,768],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.add","A":[{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_2/MultiHeadAttention/Linear/"},"N":"struct_name"}],"I":[{"%":333},{"%":163}],"O":[{"%":334,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512,768],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.full_int_array","A":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_i64","D":0},{"#":"0.a_i64","D":0},{"#":"0.a_i64","D":12},{"#":"0.a_i64","D":64}]},"N":"value"},{"AT":{"#":"1.a_dtype","D":"int64"},"N":"dtype"},{"AT":{"#":"1.a_place","D":[1,0,""]},"N":"place"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_2/MultiHeadAttention/"},"N":"struct_name"}],"I":[],"O":[{"%":335,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_i64"},[4],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":true}]},"N":"stop_gradient"}]},{"#":"1.reshape","A":[{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_2/MultiHeadAttention/"},"N":"struct_name"}],"I":[{"%":334},{"%":335}],"O":[{"%":336,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512,12,64],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.transpose","A":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_i32","D":0},{"#":"0.a_i32","D":2},{"#":"0.a_i32","D":1},{"#":"0.a_i32","D":3}]},"N":"perm"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_2/MultiHeadAttention/"},"N":"struct_name"}],"I":[{"%":336}],"O":[{"%":337,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,12,512,64],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.matmul","A":[{"AT":{"#":"0.a_bool","D":false},"N":"transpose_x"},{"AT":{"#":"0.a_bool","D":false},"N":"transpose_y"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_2/MultiHeadAttention/Linear_1/"},"N":"struct_name"}],"I":[{"%":330},{"%":162}],"O":[{"%":338,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512,768],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.add","A":[{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_2/MultiHeadAttention/Linear_1/"},"N":"struct_name"}],"I":[{"%":338},{"%":161}],"O":[{"%":339,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512,768],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.matmul","A":[{"AT":{"#":"0.a_bool","D":false},"N":"transpose_x"},{"AT":{"#":"0.a_bool","D":false},"N":"transpose_y"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_2/MultiHeadAttention/Linear_2/"},"N":"struct_name"}],"I":[{"%":330},{"%":160}],"O":[{"%":340,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512,768],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.add","A":[{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_2/MultiHeadAttention/Linear_2/"},"N":"struct_name"}],"I":[{"%":340},{"%":159}],"O":[{"%":341,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512,768],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.full_int_array","A":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_i64","D":0},{"#":"0.a_i64","D":0},{"#":"0.a_i64","D":12},{"#":"0.a_i64","D":64}]},"N":"value"},{"AT":{"#":"1.a_dtype","D":"int64"},"N":"dtype"},{"AT":{"#":"1.a_place","D":[1,0,""]},"N":"place"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_2/MultiHeadAttention/"},"N":"struct_name"}],"I":[],"O":[{"%":342,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_i64"},[4],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":true}]},"N":"stop_gradient"}]},{"#":"1.reshape","A":[{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_2/MultiHeadAttention/"},"N":"struct_name"}],"I":[{"%":339},{"%":342}],"O":[{"%":343,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512,12,64],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.transpose","A":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_i32","D":0},{"#":"0.a_i32","D":2},{"#":"0.a_i32","D":1},{"#":"0.a_i32","D":3}]},"N":"perm"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_2/MultiHeadAttention/"},"N":"struct_name"}],"I":[{"%":343}],"O":[{"%":344,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,12,512,64],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.full_int_array","A":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_i64","D":0},{"#":"0.a_i64","D":0},{"#":"0.a_i64","D":12},{"#":"0.a_i64","D":64}]},"N":"value"},{"AT":{"#":"1.a_dtype","D":"int64"},"N":"dtype"},{"AT":{"#":"1.a_place","D":[1,0,""]},"N":"place"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_2/MultiHeadAttention/"},"N":"struct_name"}],"I":[],"O":[{"%":345,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_i64"},[4],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":true}]},"N":"stop_gradient"}]},{"#":"1.reshape","A":[{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_2/MultiHeadAttention/"},"N":"struct_name"}],"I":[{"%":341},{"%":345}],"O":[{"%":346,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512,12,64],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.transpose","A":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_i32","D":0},{"#":"0.a_i32","D":2},{"#":"0.a_i32","D":1},{"#":"0.a_i32","D":3}]},"N":"perm"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_2/MultiHeadAttention/"},"N":"struct_name"}],"I":[{"%":346}],"O":[{"%":347,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,12,512,64],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.full","A":[{"AT":{"#":"1.a_intarray","D":[1]},"N":"shape"},{"AT":{"#":"0.a_f64","D":0.125},"N":"value"},{"AT":{"#":"1.a_dtype","D":"float32"},"N":"dtype"},{"AT":{"#":"1.a_place","D":[1,0,""]},"N":"place"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_2/MultiHeadAttention/"},"N":"struct_name"}],"I":[],"O":[{"%":348,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[1],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":true}]},"N":"stop_gradient"}]},{"#":"1.scale","A":[{"AT":{"#":"0.a_f32","D":0.0},"N":"bias"},{"AT":{"#":"0.a_bool","D":true},"N":"bias_after_scale"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_2/MultiHeadAttention/"},"N":"struct_name"}],"I":[{"%":337},{"%":348}],"O":[{"%":349,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,12,512,64],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.matmul","A":[{"AT":{"#":"0.a_bool","D":false},"N":"transpose_x"},{"AT":{"#":"0.a_bool","D":true},"N":"transpose_y"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_2/MultiHeadAttention/"},"N":"struct_name"}],"I":[{"%":349},{"%":344}],"O":[{"%":350,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,12,512,512],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.add","A":[{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_2/MultiHeadAttention/"},"N":"struct_name"}],"I":[{"%":350},{"%":212}],"O":[{"%":351,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,12,512,512],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.softmax","A":[{"AT":{"#":"0.a_i32","D":-1},"N":"axis"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_2/MultiHeadAttention/"},"N":"struct_name"}],"I":[{"%":351}],"O":[{"%":352,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,12,512,512],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.full","A":[{"AT":{"#":"1.a_intarray","D":[1]},"N":"shape"},{"AT":{"#":"0.a_f64","D":0.10000000149011612},"N":"value"},{"AT":{"#":"1.a_dtype","D":"float32"},"N":"dtype"},{"AT":{"#":"1.a_place","D":[1,0,""]},"N":"place"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_2/MultiHeadAttention/"},"N":"struct_name"}],"I":[],"O":[{"%":353,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[1],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":true}]},"N":"stop_gradient"}]},{"#":"1.dropout","A":[{"AT":{"#":"0.a_bool","D":true},"N":"is_test"},{"AT":{"#":"0.a_str","D":"upscale_in_train"},"N":"mode"},{"AT":{"#":"0.a_i32","D":0},"N":"seed"},{"AT":{"#":"0.a_bool","D":false},"N":"fix_seed"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_2/MultiHeadAttention/"},"N":"struct_name"}],"I":[{"%":352},{"%":0},{"%":353}],"O":[{"%":354,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,12,512,512],"NCHW",[],0]}},{"%":355,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_ui8"},[-1,12,512,512],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false},{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.matmul","A":[{"AT":{"#":"0.a_bool","D":false},"N":"transpose_x"},{"AT":{"#":"0.a_bool","D":false},"N":"transpose_y"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_2/MultiHeadAttention/"},"N":"struct_name"}],"I":[{"%":354},{"%":347}],"O":[{"%":356,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,12,512,64],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.transpose","A":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_i32","D":0},{"#":"0.a_i32","D":2},{"#":"0.a_i32","D":1},{"#":"0.a_i32","D":3}]},"N":"perm"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_2/MultiHeadAttention/"},"N":"struct_name"}],"I":[{"%":356}],"O":[{"%":357,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512,12,64],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.full_int_array","A":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_i64","D":0},{"#":"0.a_i64","D":0},{"#":"0.a_i64","D":768}]},"N":"value"},{"AT":{"#":"1.a_dtype","D":"int64"},"N":"dtype"},{"AT":{"#":"1.a_place","D":[1,0,""]},"N":"place"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_2/MultiHeadAttention/"},"N":"struct_name"}],"I":[],"O":[{"%":358,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_i64"},[3],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":true}]},"N":"stop_gradient"}]},{"#":"1.reshape","A":[{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_2/MultiHeadAttention/"},"N":"struct_name"}],"I":[{"%":357},{"%":358}],"O":[{"%":359,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512,768],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.matmul","A":[{"AT":{"#":"0.a_bool","D":false},"N":"transpose_x"},{"AT":{"#":"0.a_bool","D":false},"N":"transpose_y"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_2/MultiHeadAttention/Linear_3/"},"N":"struct_name"}],"I":[{"%":359},{"%":158}],"O":[{"%":360,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512,768],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.add","A":[{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_2/MultiHeadAttention/Linear_3/"},"N":"struct_name"}],"I":[{"%":360},{"%":157}],"O":[{"%":361,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512,768],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.full","A":[{"AT":{"#":"1.a_intarray","D":[1]},"N":"shape"},{"AT":{"#":"0.a_f64","D":0.10000000149011612},"N":"value"},{"AT":{"#":"1.a_dtype","D":"float32"},"N":"dtype"},{"AT":{"#":"1.a_place","D":[1,0,""]},"N":"place"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_2/Dropout/"},"N":"struct_name"}],"I":[],"O":[{"%":362,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[1],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":true}]},"N":"stop_gradient"}]},{"#":"1.dropout","A":[{"AT":{"#":"0.a_bool","D":true},"N":"is_test"},{"AT":{"#":"0.a_str","D":"upscale_in_train"},"N":"mode"},{"AT":{"#":"0.a_i32","D":0},"N":"seed"},{"AT":{"#":"0.a_bool","D":false},"N":"fix_seed"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_2/Dropout/"},"N":"struct_name"}],"I":[{"%":361},{"%":0},{"%":362}],"O":[{"%":363,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512,768],"NCHW",[],0]}},{"%":364,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_ui8"},[-1,512,768],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false},{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.add","A":[{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_2/"},"N":"struct_name"}],"I":[{"%":330},{"%":363}],"O":[{"%":365,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512,768],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.layer_norm","A":[{"AT":{"#":"0.a_f32","D":9.999999960041972e-13},"N":"epsilon"},{"AT":{"#":"0.a_i32","D":2},"N":"begin_norm_axis"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_2/LayerNorm/"},"N":"struct_name"}],"I":[{"%":365},{"%":152},{"%":151}],"O":[{"%":366,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512,768],"NCHW",[],0]}},{"%":367,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512],"NCHW",[],0]}},{"%":368,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false},{"#":"0.a_bool","D":false},{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.matmul","A":[{"AT":{"#":"0.a_bool","D":false},"N":"transpose_x"},{"AT":{"#":"0.a_bool","D":false},"N":"transpose_y"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_2/Linear/"},"N":"struct_name"}],"I":[{"%":366},{"%":156}],"O":[{"%":369,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512,3072],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.add","A":[{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_2/Linear/"},"N":"struct_name"}],"I":[{"%":369},{"%":155}],"O":[{"%":370,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512,3072],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.gelu","A":[{"AT":{"#":"0.a_bool","D":false},"N":"approximate"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_2/"},"N":"struct_name"}],"I":[{"%":370}],"O":[{"%":371,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512,3072],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.matmul","A":[{"AT":{"#":"0.a_bool","D":false},"N":"transpose_x"},{"AT":{"#":"0.a_bool","D":false},"N":"transpose_y"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_2/Linear_1/"},"N":"struct_name"}],"I":[{"%":371},{"%":154}],"O":[{"%":372,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512,768],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.add","A":[{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_2/Linear_1/"},"N":"struct_name"}],"I":[{"%":372},{"%":153}],"O":[{"%":373,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512,768],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.full","A":[{"AT":{"#":"1.a_intarray","D":[1]},"N":"shape"},{"AT":{"#":"0.a_f64","D":0.10000000149011612},"N":"value"},{"AT":{"#":"1.a_dtype","D":"float32"},"N":"dtype"},{"AT":{"#":"1.a_place","D":[1,0,""]},"N":"place"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_2/Dropout_2/"},"N":"struct_name"}],"I":[],"O":[{"%":374,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[1],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":true}]},"N":"stop_gradient"}]},{"#":"1.dropout","A":[{"AT":{"#":"0.a_bool","D":true},"N":"is_test"},{"AT":{"#":"0.a_str","D":"upscale_in_train"},"N":"mode"},{"AT":{"#":"0.a_i32","D":0},"N":"seed"},{"AT":{"#":"0.a_bool","D":false},"N":"fix_seed"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_2/Dropout_2/"},"N":"struct_name"}],"I":[{"%":373},{"%":0},{"%":374}],"O":[{"%":375,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512,768],"NCHW",[],0]}},{"%":376,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_ui8"},[-1,512,768],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false},{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.add","A":[{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_2/"},"N":"struct_name"}],"I":[{"%":366},{"%":375}],"O":[{"%":377,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512,768],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.layer_norm","A":[{"AT":{"#":"0.a_f32","D":9.999999960041972e-13},"N":"epsilon"},{"AT":{"#":"0.a_i32","D":2},"N":"begin_norm_axis"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_2/LayerNorm_1/"},"N":"struct_name"}],"I":[{"%":377},{"%":150},{"%":149}],"O":[{"%":378,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512,768],"NCHW",[],0]}},{"%":379,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512],"NCHW",[],0]}},{"%":380,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false},{"#":"0.a_bool","D":false},{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.matmul","A":[{"AT":{"#":"0.a_bool","D":false},"N":"transpose_x"},{"AT":{"#":"0.a_bool","D":false},"N":"transpose_y"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_3/MultiHeadAttention/Linear/"},"N":"struct_name"}],"I":[{"%":378},{"%":148}],"O":[{"%":381,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512,768],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.add","A":[{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_3/MultiHeadAttention/Linear/"},"N":"struct_name"}],"I":[{"%":381},{"%":147}],"O":[{"%":382,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512,768],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.full_int_array","A":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_i64","D":0},{"#":"0.a_i64","D":0},{"#":"0.a_i64","D":12},{"#":"0.a_i64","D":64}]},"N":"value"},{"AT":{"#":"1.a_dtype","D":"int64"},"N":"dtype"},{"AT":{"#":"1.a_place","D":[1,0,""]},"N":"place"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_3/MultiHeadAttention/"},"N":"struct_name"}],"I":[],"O":[{"%":383,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_i64"},[4],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":true}]},"N":"stop_gradient"}]},{"#":"1.reshape","A":[{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_3/MultiHeadAttention/"},"N":"struct_name"}],"I":[{"%":382},{"%":383}],"O":[{"%":384,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512,12,64],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.transpose","A":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_i32","D":0},{"#":"0.a_i32","D":2},{"#":"0.a_i32","D":1},{"#":"0.a_i32","D":3}]},"N":"perm"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_3/MultiHeadAttention/"},"N":"struct_name"}],"I":[{"%":384}],"O":[{"%":385,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,12,512,64],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.matmul","A":[{"AT":{"#":"0.a_bool","D":false},"N":"transpose_x"},{"AT":{"#":"0.a_bool","D":false},"N":"transpose_y"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_3/MultiHeadAttention/Linear_1/"},"N":"struct_name"}],"I":[{"%":378},{"%":146}],"O":[{"%":386,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512,768],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.add","A":[{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_3/MultiHeadAttention/Linear_1/"},"N":"struct_name"}],"I":[{"%":386},{"%":145}],"O":[{"%":387,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512,768],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.matmul","A":[{"AT":{"#":"0.a_bool","D":false},"N":"transpose_x"},{"AT":{"#":"0.a_bool","D":false},"N":"transpose_y"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_3/MultiHeadAttention/Linear_2/"},"N":"struct_name"}],"I":[{"%":378},{"%":144}],"O":[{"%":388,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512,768],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.add","A":[{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_3/MultiHeadAttention/Linear_2/"},"N":"struct_name"}],"I":[{"%":388},{"%":143}],"O":[{"%":389,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512,768],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.full_int_array","A":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_i64","D":0},{"#":"0.a_i64","D":0},{"#":"0.a_i64","D":12},{"#":"0.a_i64","D":64}]},"N":"value"},{"AT":{"#":"1.a_dtype","D":"int64"},"N":"dtype"},{"AT":{"#":"1.a_place","D":[1,0,""]},"N":"place"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_3/MultiHeadAttention/"},"N":"struct_name"}],"I":[],"O":[{"%":390,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_i64"},[4],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":true}]},"N":"stop_gradient"}]},{"#":"1.reshape","A":[{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_3/MultiHeadAttention/"},"N":"struct_name"}],"I":[{"%":387},{"%":390}],"O":[{"%":391,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512,12,64],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.transpose","A":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_i32","D":0},{"#":"0.a_i32","D":2},{"#":"0.a_i32","D":1},{"#":"0.a_i32","D":3}]},"N":"perm"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_3/MultiHeadAttention/"},"N":"struct_name"}],"I":[{"%":391}],"O":[{"%":392,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,12,512,64],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.full_int_array","A":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_i64","D":0},{"#":"0.a_i64","D":0},{"#":"0.a_i64","D":12},{"#":"0.a_i64","D":64}]},"N":"value"},{"AT":{"#":"1.a_dtype","D":"int64"},"N":"dtype"},{"AT":{"#":"1.a_place","D":[1,0,""]},"N":"place"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_3/MultiHeadAttention/"},"N":"struct_name"}],"I":[],"O":[{"%":393,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_i64"},[4],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":true}]},"N":"stop_gradient"}]},{"#":"1.reshape","A":[{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_3/MultiHeadAttention/"},"N":"struct_name"}],"I":[{"%":389},{"%":393}],"O":[{"%":394,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512,12,64],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.transpose","A":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_i32","D":0},{"#":"0.a_i32","D":2},{"#":"0.a_i32","D":1},{"#":"0.a_i32","D":3}]},"N":"perm"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_3/MultiHeadAttention/"},"N":"struct_name"}],"I":[{"%":394}],"O":[{"%":395,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,12,512,64],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.full","A":[{"AT":{"#":"1.a_intarray","D":[1]},"N":"shape"},{"AT":{"#":"0.a_f64","D":0.125},"N":"value"},{"AT":{"#":"1.a_dtype","D":"float32"},"N":"dtype"},{"AT":{"#":"1.a_place","D":[1,0,""]},"N":"place"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_3/MultiHeadAttention/"},"N":"struct_name"}],"I":[],"O":[{"%":396,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[1],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":true}]},"N":"stop_gradient"}]},{"#":"1.scale","A":[{"AT":{"#":"0.a_f32","D":0.0},"N":"bias"},{"AT":{"#":"0.a_bool","D":true},"N":"bias_after_scale"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_3/MultiHeadAttention/"},"N":"struct_name"}],"I":[{"%":385},{"%":396}],"O":[{"%":397,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,12,512,64],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.matmul","A":[{"AT":{"#":"0.a_bool","D":false},"N":"transpose_x"},{"AT":{"#":"0.a_bool","D":true},"N":"transpose_y"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_3/MultiHeadAttention/"},"N":"struct_name"}],"I":[{"%":397},{"%":392}],"O":[{"%":398,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,12,512,512],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.add","A":[{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_3/MultiHeadAttention/"},"N":"struct_name"}],"I":[{"%":398},{"%":212}],"O":[{"%":399,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,12,512,512],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.softmax","A":[{"AT":{"#":"0.a_i32","D":-1},"N":"axis"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_3/MultiHeadAttention/"},"N":"struct_name"}],"I":[{"%":399}],"O":[{"%":400,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,12,512,512],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.full","A":[{"AT":{"#":"1.a_intarray","D":[1]},"N":"shape"},{"AT":{"#":"0.a_f64","D":0.10000000149011612},"N":"value"},{"AT":{"#":"1.a_dtype","D":"float32"},"N":"dtype"},{"AT":{"#":"1.a_place","D":[1,0,""]},"N":"place"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_3/MultiHeadAttention/"},"N":"struct_name"}],"I":[],"O":[{"%":401,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[1],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":true}]},"N":"stop_gradient"}]},{"#":"1.dropout","A":[{"AT":{"#":"0.a_bool","D":true},"N":"is_test"},{"AT":{"#":"0.a_str","D":"upscale_in_train"},"N":"mode"},{"AT":{"#":"0.a_i32","D":0},"N":"seed"},{"AT":{"#":"0.a_bool","D":false},"N":"fix_seed"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_3/MultiHeadAttention/"},"N":"struct_name"}],"I":[{"%":400},{"%":0},{"%":401}],"O":[{"%":402,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,12,512,512],"NCHW",[],0]}},{"%":403,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_ui8"},[-1,12,512,512],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false},{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.matmul","A":[{"AT":{"#":"0.a_bool","D":false},"N":"transpose_x"},{"AT":{"#":"0.a_bool","D":false},"N":"transpose_y"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_3/MultiHeadAttention/"},"N":"struct_name"}],"I":[{"%":402},{"%":395}],"O":[{"%":404,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,12,512,64],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.transpose","A":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_i32","D":0},{"#":"0.a_i32","D":2},{"#":"0.a_i32","D":1},{"#":"0.a_i32","D":3}]},"N":"perm"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_3/MultiHeadAttention/"},"N":"struct_name"}],"I":[{"%":404}],"O":[{"%":405,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512,12,64],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.full_int_array","A":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_i64","D":0},{"#":"0.a_i64","D":0},{"#":"0.a_i64","D":768}]},"N":"value"},{"AT":{"#":"1.a_dtype","D":"int64"},"N":"dtype"},{"AT":{"#":"1.a_place","D":[1,0,""]},"N":"place"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_3/MultiHeadAttention/"},"N":"struct_name"}],"I":[],"O":[{"%":406,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_i64"},[3],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":true}]},"N":"stop_gradient"}]},{"#":"1.reshape","A":[{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_3/MultiHeadAttention/"},"N":"struct_name"}],"I":[{"%":405},{"%":406}],"O":[{"%":407,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512,768],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.matmul","A":[{"AT":{"#":"0.a_bool","D":false},"N":"transpose_x"},{"AT":{"#":"0.a_bool","D":false},"N":"transpose_y"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_3/MultiHeadAttention/Linear_3/"},"N":"struct_name"}],"I":[{"%":407},{"%":142}],"O":[{"%":408,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512,768],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.add","A":[{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_3/MultiHeadAttention/Linear_3/"},"N":"struct_name"}],"I":[{"%":408},{"%":141}],"O":[{"%":409,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512,768],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.full","A":[{"AT":{"#":"1.a_intarray","D":[1]},"N":"shape"},{"AT":{"#":"0.a_f64","D":0.10000000149011612},"N":"value"},{"AT":{"#":"1.a_dtype","D":"float32"},"N":"dtype"},{"AT":{"#":"1.a_place","D":[1,0,""]},"N":"place"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_3/Dropout/"},"N":"struct_name"}],"I":[],"O":[{"%":410,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[1],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":true}]},"N":"stop_gradient"}]},{"#":"1.dropout","A":[{"AT":{"#":"0.a_bool","D":true},"N":"is_test"},{"AT":{"#":"0.a_str","D":"upscale_in_train"},"N":"mode"},{"AT":{"#":"0.a_i32","D":0},"N":"seed"},{"AT":{"#":"0.a_bool","D":false},"N":"fix_seed"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_3/Dropout/"},"N":"struct_name"}],"I":[{"%":409},{"%":0},{"%":410}],"O":[{"%":411,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512,768],"NCHW",[],0]}},{"%":412,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_ui8"},[-1,512,768],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false},{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.add","A":[{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_3/"},"N":"struct_name"}],"I":[{"%":378},{"%":411}],"O":[{"%":413,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512,768],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.layer_norm","A":[{"AT":{"#":"0.a_f32","D":9.999999960041972e-13},"N":"epsilon"},{"AT":{"#":"0.a_i32","D":2},"N":"begin_norm_axis"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_3/LayerNorm/"},"N":"struct_name"}],"I":[{"%":413},{"%":136},{"%":135}],"O":[{"%":414,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512,768],"NCHW",[],0]}},{"%":415,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512],"NCHW",[],0]}},{"%":416,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false},{"#":"0.a_bool","D":false},{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.matmul","A":[{"AT":{"#":"0.a_bool","D":false},"N":"transpose_x"},{"AT":{"#":"0.a_bool","D":false},"N":"transpose_y"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_3/Linear/"},"N":"struct_name"}],"I":[{"%":414},{"%":140}],"O":[{"%":417,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512,3072],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.add","A":[{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_3/Linear/"},"N":"struct_name"}],"I":[{"%":417},{"%":139}],"O":[{"%":418,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512,3072],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.gelu","A":[{"AT":{"#":"0.a_bool","D":false},"N":"approximate"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_3/"},"N":"struct_name"}],"I":[{"%":418}],"O":[{"%":419,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512,3072],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.matmul","A":[{"AT":{"#":"0.a_bool","D":false},"N":"transpose_x"},{"AT":{"#":"0.a_bool","D":false},"N":"transpose_y"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_3/Linear_1/"},"N":"struct_name"}],"I":[{"%":419},{"%":138}],"O":[{"%":420,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512,768],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.add","A":[{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_3/Linear_1/"},"N":"struct_name"}],"I":[{"%":420},{"%":137}],"O":[{"%":421,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512,768],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.full","A":[{"AT":{"#":"1.a_intarray","D":[1]},"N":"shape"},{"AT":{"#":"0.a_f64","D":0.10000000149011612},"N":"value"},{"AT":{"#":"1.a_dtype","D":"float32"},"N":"dtype"},{"AT":{"#":"1.a_place","D":[1,0,""]},"N":"place"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_3/Dropout_2/"},"N":"struct_name"}],"I":[],"O":[{"%":422,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[1],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":true}]},"N":"stop_gradient"}]},{"#":"1.dropout","A":[{"AT":{"#":"0.a_bool","D":true},"N":"is_test"},{"AT":{"#":"0.a_str","D":"upscale_in_train"},"N":"mode"},{"AT":{"#":"0.a_i32","D":0},"N":"seed"},{"AT":{"#":"0.a_bool","D":false},"N":"fix_seed"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_3/Dropout_2/"},"N":"struct_name"}],"I":[{"%":421},{"%":0},{"%":422}],"O":[{"%":423,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512,768],"NCHW",[],0]}},{"%":424,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_ui8"},[-1,512,768],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false},{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.add","A":[{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_3/"},"N":"struct_name"}],"I":[{"%":414},{"%":423}],"O":[{"%":425,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512,768],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.layer_norm","A":[{"AT":{"#":"0.a_f32","D":9.999999960041972e-13},"N":"epsilon"},{"AT":{"#":"0.a_i32","D":2},"N":"begin_norm_axis"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_3/LayerNorm_1/"},"N":"struct_name"}],"I":[{"%":425},{"%":134},{"%":133}],"O":[{"%":426,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512,768],"NCHW",[],0]}},{"%":427,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512],"NCHW",[],0]}},{"%":428,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false},{"#":"0.a_bool","D":false},{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.matmul","A":[{"AT":{"#":"0.a_bool","D":false},"N":"transpose_x"},{"AT":{"#":"0.a_bool","D":false},"N":"transpose_y"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_4/MultiHeadAttention/Linear/"},"N":"struct_name"}],"I":[{"%":426},{"%":132}],"O":[{"%":429,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512,768],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.add","A":[{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_4/MultiHeadAttention/Linear/"},"N":"struct_name"}],"I":[{"%":429},{"%":131}],"O":[{"%":430,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512,768],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.full_int_array","A":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_i64","D":0},{"#":"0.a_i64","D":0},{"#":"0.a_i64","D":12},{"#":"0.a_i64","D":64}]},"N":"value"},{"AT":{"#":"1.a_dtype","D":"int64"},"N":"dtype"},{"AT":{"#":"1.a_place","D":[1,0,""]},"N":"place"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_4/MultiHeadAttention/"},"N":"struct_name"}],"I":[],"O":[{"%":431,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_i64"},[4],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":true}]},"N":"stop_gradient"}]},{"#":"1.reshape","A":[{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_4/MultiHeadAttention/"},"N":"struct_name"}],"I":[{"%":430},{"%":431}],"O":[{"%":432,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512,12,64],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.transpose","A":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_i32","D":0},{"#":"0.a_i32","D":2},{"#":"0.a_i32","D":1},{"#":"0.a_i32","D":3}]},"N":"perm"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_4/MultiHeadAttention/"},"N":"struct_name"}],"I":[{"%":432}],"O":[{"%":433,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,12,512,64],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.matmul","A":[{"AT":{"#":"0.a_bool","D":false},"N":"transpose_x"},{"AT":{"#":"0.a_bool","D":false},"N":"transpose_y"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_4/MultiHeadAttention/Linear_1/"},"N":"struct_name"}],"I":[{"%":426},{"%":130}],"O":[{"%":434,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512,768],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.add","A":[{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_4/MultiHeadAttention/Linear_1/"},"N":"struct_name"}],"I":[{"%":434},{"%":129}],"O":[{"%":435,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512,768],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.matmul","A":[{"AT":{"#":"0.a_bool","D":false},"N":"transpose_x"},{"AT":{"#":"0.a_bool","D":false},"N":"transpose_y"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_4/MultiHeadAttention/Linear_2/"},"N":"struct_name"}],"I":[{"%":426},{"%":128}],"O":[{"%":436,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512,768],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.add","A":[{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_4/MultiHeadAttention/Linear_2/"},"N":"struct_name"}],"I":[{"%":436},{"%":127}],"O":[{"%":437,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512,768],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.full_int_array","A":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_i64","D":0},{"#":"0.a_i64","D":0},{"#":"0.a_i64","D":12},{"#":"0.a_i64","D":64}]},"N":"value"},{"AT":{"#":"1.a_dtype","D":"int64"},"N":"dtype"},{"AT":{"#":"1.a_place","D":[1,0,""]},"N":"place"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_4/MultiHeadAttention/"},"N":"struct_name"}],"I":[],"O":[{"%":438,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_i64"},[4],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":true}]},"N":"stop_gradient"}]},{"#":"1.reshape","A":[{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_4/MultiHeadAttention/"},"N":"struct_name"}],"I":[{"%":435},{"%":438}],"O":[{"%":439,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512,12,64],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.transpose","A":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_i32","D":0},{"#":"0.a_i32","D":2},{"#":"0.a_i32","D":1},{"#":"0.a_i32","D":3}]},"N":"perm"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_4/MultiHeadAttention/"},"N":"struct_name"}],"I":[{"%":439}],"O":[{"%":440,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,12,512,64],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.full_int_array","A":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_i64","D":0},{"#":"0.a_i64","D":0},{"#":"0.a_i64","D":12},{"#":"0.a_i64","D":64}]},"N":"value"},{"AT":{"#":"1.a_dtype","D":"int64"},"N":"dtype"},{"AT":{"#":"1.a_place","D":[1,0,""]},"N":"place"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_4/MultiHeadAttention/"},"N":"struct_name"}],"I":[],"O":[{"%":441,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_i64"},[4],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":true}]},"N":"stop_gradient"}]},{"#":"1.reshape","A":[{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_4/MultiHeadAttention/"},"N":"struct_name"}],"I":[{"%":437},{"%":441}],"O":[{"%":442,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512,12,64],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.transpose","A":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_i32","D":0},{"#":"0.a_i32","D":2},{"#":"0.a_i32","D":1},{"#":"0.a_i32","D":3}]},"N":"perm"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_4/MultiHeadAttention/"},"N":"struct_name"}],"I":[{"%":442}],"O":[{"%":443,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,12,512,64],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.full","A":[{"AT":{"#":"1.a_intarray","D":[1]},"N":"shape"},{"AT":{"#":"0.a_f64","D":0.125},"N":"value"},{"AT":{"#":"1.a_dtype","D":"float32"},"N":"dtype"},{"AT":{"#":"1.a_place","D":[1,0,""]},"N":"place"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_4/MultiHeadAttention/"},"N":"struct_name"}],"I":[],"O":[{"%":444,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[1],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":true}]},"N":"stop_gradient"}]},{"#":"1.scale","A":[{"AT":{"#":"0.a_f32","D":0.0},"N":"bias"},{"AT":{"#":"0.a_bool","D":true},"N":"bias_after_scale"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_4/MultiHeadAttention/"},"N":"struct_name"}],"I":[{"%":433},{"%":444}],"O":[{"%":445,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,12,512,64],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.matmul","A":[{"AT":{"#":"0.a_bool","D":false},"N":"transpose_x"},{"AT":{"#":"0.a_bool","D":true},"N":"transpose_y"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_4/MultiHeadAttention/"},"N":"struct_name"}],"I":[{"%":445},{"%":440}],"O":[{"%":446,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,12,512,512],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.add","A":[{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_4/MultiHeadAttention/"},"N":"struct_name"}],"I":[{"%":446},{"%":212}],"O":[{"%":447,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,12,512,512],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.softmax","A":[{"AT":{"#":"0.a_i32","D":-1},"N":"axis"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_4/MultiHeadAttention/"},"N":"struct_name"}],"I":[{"%":447}],"O":[{"%":448,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,12,512,512],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.full","A":[{"AT":{"#":"1.a_intarray","D":[1]},"N":"shape"},{"AT":{"#":"0.a_f64","D":0.10000000149011612},"N":"value"},{"AT":{"#":"1.a_dtype","D":"float32"},"N":"dtype"},{"AT":{"#":"1.a_place","D":[1,0,""]},"N":"place"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_4/MultiHeadAttention/"},"N":"struct_name"}],"I":[],"O":[{"%":449,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[1],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":true}]},"N":"stop_gradient"}]},{"#":"1.dropout","A":[{"AT":{"#":"0.a_bool","D":true},"N":"is_test"},{"AT":{"#":"0.a_str","D":"upscale_in_train"},"N":"mode"},{"AT":{"#":"0.a_i32","D":0},"N":"seed"},{"AT":{"#":"0.a_bool","D":false},"N":"fix_seed"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_4/MultiHeadAttention/"},"N":"struct_name"}],"I":[{"%":448},{"%":0},{"%":449}],"O":[{"%":450,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,12,512,512],"NCHW",[],0]}},{"%":451,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_ui8"},[-1,12,512,512],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false},{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.matmul","A":[{"AT":{"#":"0.a_bool","D":false},"N":"transpose_x"},{"AT":{"#":"0.a_bool","D":false},"N":"transpose_y"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_4/MultiHeadAttention/"},"N":"struct_name"}],"I":[{"%":450},{"%":443}],"O":[{"%":452,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,12,512,64],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.transpose","A":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_i32","D":0},{"#":"0.a_i32","D":2},{"#":"0.a_i32","D":1},{"#":"0.a_i32","D":3}]},"N":"perm"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_4/MultiHeadAttention/"},"N":"struct_name"}],"I":[{"%":452}],"O":[{"%":453,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512,12,64],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.full_int_array","A":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_i64","D":0},{"#":"0.a_i64","D":0},{"#":"0.a_i64","D":768}]},"N":"value"},{"AT":{"#":"1.a_dtype","D":"int64"},"N":"dtype"},{"AT":{"#":"1.a_place","D":[1,0,""]},"N":"place"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_4/MultiHeadAttention/"},"N":"struct_name"}],"I":[],"O":[{"%":454,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_i64"},[3],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":true}]},"N":"stop_gradient"}]},{"#":"1.reshape","A":[{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_4/MultiHeadAttention/"},"N":"struct_name"}],"I":[{"%":453},{"%":454}],"O":[{"%":455,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512,768],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.matmul","A":[{"AT":{"#":"0.a_bool","D":false},"N":"transpose_x"},{"AT":{"#":"0.a_bool","D":false},"N":"transpose_y"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_4/MultiHeadAttention/Linear_3/"},"N":"struct_name"}],"I":[{"%":455},{"%":126}],"O":[{"%":456,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512,768],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.add","A":[{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_4/MultiHeadAttention/Linear_3/"},"N":"struct_name"}],"I":[{"%":456},{"%":125}],"O":[{"%":457,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512,768],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.full","A":[{"AT":{"#":"1.a_intarray","D":[1]},"N":"shape"},{"AT":{"#":"0.a_f64","D":0.10000000149011612},"N":"value"},{"AT":{"#":"1.a_dtype","D":"float32"},"N":"dtype"},{"AT":{"#":"1.a_place","D":[1,0,""]},"N":"place"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_4/Dropout/"},"N":"struct_name"}],"I":[],"O":[{"%":458,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[1],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":true}]},"N":"stop_gradient"}]},{"#":"1.dropout","A":[{"AT":{"#":"0.a_bool","D":true},"N":"is_test"},{"AT":{"#":"0.a_str","D":"upscale_in_train"},"N":"mode"},{"AT":{"#":"0.a_i32","D":0},"N":"seed"},{"AT":{"#":"0.a_bool","D":false},"N":"fix_seed"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_4/Dropout/"},"N":"struct_name"}],"I":[{"%":457},{"%":0},{"%":458}],"O":[{"%":459,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512,768],"NCHW",[],0]}},{"%":460,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_ui8"},[-1,512,768],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false},{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.add","A":[{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_4/"},"N":"struct_name"}],"I":[{"%":426},{"%":459}],"O":[{"%":461,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512,768],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.layer_norm","A":[{"AT":{"#":"0.a_f32","D":9.999999960041972e-13},"N":"epsilon"},{"AT":{"#":"0.a_i32","D":2},"N":"begin_norm_axis"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_4/LayerNorm/"},"N":"struct_name"}],"I":[{"%":461},{"%":120},{"%":119}],"O":[{"%":462,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512,768],"NCHW",[],0]}},{"%":463,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512],"NCHW",[],0]}},{"%":464,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false},{"#":"0.a_bool","D":false},{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.matmul","A":[{"AT":{"#":"0.a_bool","D":false},"N":"transpose_x"},{"AT":{"#":"0.a_bool","D":false},"N":"transpose_y"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_4/Linear/"},"N":"struct_name"}],"I":[{"%":462},{"%":124}],"O":[{"%":465,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512,3072],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.add","A":[{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_4/Linear/"},"N":"struct_name"}],"I":[{"%":465},{"%":123}],"O":[{"%":466,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512,3072],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.gelu","A":[{"AT":{"#":"0.a_bool","D":false},"N":"approximate"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_4/"},"N":"struct_name"}],"I":[{"%":466}],"O":[{"%":467,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512,3072],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.matmul","A":[{"AT":{"#":"0.a_bool","D":false},"N":"transpose_x"},{"AT":{"#":"0.a_bool","D":false},"N":"transpose_y"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_4/Linear_1/"},"N":"struct_name"}],"I":[{"%":467},{"%":122}],"O":[{"%":468,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512,768],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.add","A":[{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_4/Linear_1/"},"N":"struct_name"}],"I":[{"%":468},{"%":121}],"O":[{"%":469,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512,768],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.full","A":[{"AT":{"#":"1.a_intarray","D":[1]},"N":"shape"},{"AT":{"#":"0.a_f64","D":0.10000000149011612},"N":"value"},{"AT":{"#":"1.a_dtype","D":"float32"},"N":"dtype"},{"AT":{"#":"1.a_place","D":[1,0,""]},"N":"place"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_4/Dropout_2/"},"N":"struct_name"}],"I":[],"O":[{"%":470,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[1],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":true}]},"N":"stop_gradient"}]},{"#":"1.dropout","A":[{"AT":{"#":"0.a_bool","D":true},"N":"is_test"},{"AT":{"#":"0.a_str","D":"upscale_in_train"},"N":"mode"},{"AT":{"#":"0.a_i32","D":0},"N":"seed"},{"AT":{"#":"0.a_bool","D":false},"N":"fix_seed"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_4/Dropout_2/"},"N":"struct_name"}],"I":[{"%":469},{"%":0},{"%":470}],"O":[{"%":471,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512,768],"NCHW",[],0]}},{"%":472,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_ui8"},[-1,512,768],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false},{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.add","A":[{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_4/"},"N":"struct_name"}],"I":[{"%":462},{"%":471}],"O":[{"%":473,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512,768],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.layer_norm","A":[{"AT":{"#":"0.a_f32","D":9.999999960041972e-13},"N":"epsilon"},{"AT":{"#":"0.a_i32","D":2},"N":"begin_norm_axis"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_4/LayerNorm_1/"},"N":"struct_name"}],"I":[{"%":473},{"%":118},{"%":117}],"O":[{"%":474,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512,768],"NCHW",[],0]}},{"%":475,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512],"NCHW",[],0]}},{"%":476,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false},{"#":"0.a_bool","D":false},{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.matmul","A":[{"AT":{"#":"0.a_bool","D":false},"N":"transpose_x"},{"AT":{"#":"0.a_bool","D":false},"N":"transpose_y"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_5/MultiHeadAttention/Linear/"},"N":"struct_name"}],"I":[{"%":474},{"%":116}],"O":[{"%":477,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512,768],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.add","A":[{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_5/MultiHeadAttention/Linear/"},"N":"struct_name"}],"I":[{"%":477},{"%":115}],"O":[{"%":478,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512,768],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.full_int_array","A":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_i64","D":0},{"#":"0.a_i64","D":0},{"#":"0.a_i64","D":12},{"#":"0.a_i64","D":64}]},"N":"value"},{"AT":{"#":"1.a_dtype","D":"int64"},"N":"dtype"},{"AT":{"#":"1.a_place","D":[1,0,""]},"N":"place"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_5/MultiHeadAttention/"},"N":"struct_name"}],"I":[],"O":[{"%":479,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_i64"},[4],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":true}]},"N":"stop_gradient"}]},{"#":"1.reshape","A":[{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_5/MultiHeadAttention/"},"N":"struct_name"}],"I":[{"%":478},{"%":479}],"O":[{"%":480,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512,12,64],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.transpose","A":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_i32","D":0},{"#":"0.a_i32","D":2},{"#":"0.a_i32","D":1},{"#":"0.a_i32","D":3}]},"N":"perm"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_5/MultiHeadAttention/"},"N":"struct_name"}],"I":[{"%":480}],"O":[{"%":481,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,12,512,64],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.matmul","A":[{"AT":{"#":"0.a_bool","D":false},"N":"transpose_x"},{"AT":{"#":"0.a_bool","D":false},"N":"transpose_y"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_5/MultiHeadAttention/Linear_1/"},"N":"struct_name"}],"I":[{"%":474},{"%":114}],"O":[{"%":482,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512,768],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.add","A":[{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_5/MultiHeadAttention/Linear_1/"},"N":"struct_name"}],"I":[{"%":482},{"%":113}],"O":[{"%":483,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512,768],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.matmul","A":[{"AT":{"#":"0.a_bool","D":false},"N":"transpose_x"},{"AT":{"#":"0.a_bool","D":false},"N":"transpose_y"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_5/MultiHeadAttention/Linear_2/"},"N":"struct_name"}],"I":[{"%":474},{"%":112}],"O":[{"%":484,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512,768],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.add","A":[{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_5/MultiHeadAttention/Linear_2/"},"N":"struct_name"}],"I":[{"%":484},{"%":111}],"O":[{"%":485,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512,768],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.full_int_array","A":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_i64","D":0},{"#":"0.a_i64","D":0},{"#":"0.a_i64","D":12},{"#":"0.a_i64","D":64}]},"N":"value"},{"AT":{"#":"1.a_dtype","D":"int64"},"N":"dtype"},{"AT":{"#":"1.a_place","D":[1,0,""]},"N":"place"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_5/MultiHeadAttention/"},"N":"struct_name"}],"I":[],"O":[{"%":486,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_i64"},[4],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":true}]},"N":"stop_gradient"}]},{"#":"1.reshape","A":[{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_5/MultiHeadAttention/"},"N":"struct_name"}],"I":[{"%":483},{"%":486}],"O":[{"%":487,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512,12,64],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.transpose","A":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_i32","D":0},{"#":"0.a_i32","D":2},{"#":"0.a_i32","D":1},{"#":"0.a_i32","D":3}]},"N":"perm"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_5/MultiHeadAttention/"},"N":"struct_name"}],"I":[{"%":487}],"O":[{"%":488,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,12,512,64],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.full_int_array","A":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_i64","D":0},{"#":"0.a_i64","D":0},{"#":"0.a_i64","D":12},{"#":"0.a_i64","D":64}]},"N":"value"},{"AT":{"#":"1.a_dtype","D":"int64"},"N":"dtype"},{"AT":{"#":"1.a_place","D":[1,0,""]},"N":"place"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_5/MultiHeadAttention/"},"N":"struct_name"}],"I":[],"O":[{"%":489,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_i64"},[4],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":true}]},"N":"stop_gradient"}]},{"#":"1.reshape","A":[{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_5/MultiHeadAttention/"},"N":"struct_name"}],"I":[{"%":485},{"%":489}],"O":[{"%":490,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512,12,64],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.transpose","A":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_i32","D":0},{"#":"0.a_i32","D":2},{"#":"0.a_i32","D":1},{"#":"0.a_i32","D":3}]},"N":"perm"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_5/MultiHeadAttention/"},"N":"struct_name"}],"I":[{"%":490}],"O":[{"%":491,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,12,512,64],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.full","A":[{"AT":{"#":"1.a_intarray","D":[1]},"N":"shape"},{"AT":{"#":"0.a_f64","D":0.125},"N":"value"},{"AT":{"#":"1.a_dtype","D":"float32"},"N":"dtype"},{"AT":{"#":"1.a_place","D":[1,0,""]},"N":"place"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_5/MultiHeadAttention/"},"N":"struct_name"}],"I":[],"O":[{"%":492,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[1],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":true}]},"N":"stop_gradient"}]},{"#":"1.scale","A":[{"AT":{"#":"0.a_f32","D":0.0},"N":"bias"},{"AT":{"#":"0.a_bool","D":true},"N":"bias_after_scale"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_5/MultiHeadAttention/"},"N":"struct_name"}],"I":[{"%":481},{"%":492}],"O":[{"%":493,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,12,512,64],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.matmul","A":[{"AT":{"#":"0.a_bool","D":false},"N":"transpose_x"},{"AT":{"#":"0.a_bool","D":true},"N":"transpose_y"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_5/MultiHeadAttention/"},"N":"struct_name"}],"I":[{"%":493},{"%":488}],"O":[{"%":494,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,12,512,512],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.add","A":[{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_5/MultiHeadAttention/"},"N":"struct_name"}],"I":[{"%":494},{"%":212}],"O":[{"%":495,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,12,512,512],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.softmax","A":[{"AT":{"#":"0.a_i32","D":-1},"N":"axis"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_5/MultiHeadAttention/"},"N":"struct_name"}],"I":[{"%":495}],"O":[{"%":496,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,12,512,512],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.full","A":[{"AT":{"#":"1.a_intarray","D":[1]},"N":"shape"},{"AT":{"#":"0.a_f64","D":0.10000000149011612},"N":"value"},{"AT":{"#":"1.a_dtype","D":"float32"},"N":"dtype"},{"AT":{"#":"1.a_place","D":[1,0,""]},"N":"place"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_5/MultiHeadAttention/"},"N":"struct_name"}],"I":[],"O":[{"%":497,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[1],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":true}]},"N":"stop_gradient"}]},{"#":"1.dropout","A":[{"AT":{"#":"0.a_bool","D":true},"N":"is_test"},{"AT":{"#":"0.a_str","D":"upscale_in_train"},"N":"mode"},{"AT":{"#":"0.a_i32","D":0},"N":"seed"},{"AT":{"#":"0.a_bool","D":false},"N":"fix_seed"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_5/MultiHeadAttention/"},"N":"struct_name"}],"I":[{"%":496},{"%":0},{"%":497}],"O":[{"%":498,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,12,512,512],"NCHW",[],0]}},{"%":499,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_ui8"},[-1,12,512,512],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false},{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.matmul","A":[{"AT":{"#":"0.a_bool","D":false},"N":"transpose_x"},{"AT":{"#":"0.a_bool","D":false},"N":"transpose_y"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_5/MultiHeadAttention/"},"N":"struct_name"}],"I":[{"%":498},{"%":491}],"O":[{"%":500,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,12,512,64],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.transpose","A":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_i32","D":0},{"#":"0.a_i32","D":2},{"#":"0.a_i32","D":1},{"#":"0.a_i32","D":3}]},"N":"perm"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_5/MultiHeadAttention/"},"N":"struct_name"}],"I":[{"%":500}],"O":[{"%":501,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512,12,64],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.full_int_array","A":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_i64","D":0},{"#":"0.a_i64","D":0},{"#":"0.a_i64","D":768}]},"N":"value"},{"AT":{"#":"1.a_dtype","D":"int64"},"N":"dtype"},{"AT":{"#":"1.a_place","D":[1,0,""]},"N":"place"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_5/MultiHeadAttention/"},"N":"struct_name"}],"I":[],"O":[{"%":502,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_i64"},[3],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":true}]},"N":"stop_gradient"}]},{"#":"1.reshape","A":[{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_5/MultiHeadAttention/"},"N":"struct_name"}],"I":[{"%":501},{"%":502}],"O":[{"%":503,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512,768],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.matmul","A":[{"AT":{"#":"0.a_bool","D":false},"N":"transpose_x"},{"AT":{"#":"0.a_bool","D":false},"N":"transpose_y"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_5/MultiHeadAttention/Linear_3/"},"N":"struct_name"}],"I":[{"%":503},{"%":110}],"O":[{"%":504,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512,768],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.add","A":[{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_5/MultiHeadAttention/Linear_3/"},"N":"struct_name"}],"I":[{"%":504},{"%":109}],"O":[{"%":505,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512,768],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.full","A":[{"AT":{"#":"1.a_intarray","D":[1]},"N":"shape"},{"AT":{"#":"0.a_f64","D":0.10000000149011612},"N":"value"},{"AT":{"#":"1.a_dtype","D":"float32"},"N":"dtype"},{"AT":{"#":"1.a_place","D":[1,0,""]},"N":"place"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_5/Dropout/"},"N":"struct_name"}],"I":[],"O":[{"%":506,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[1],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":true}]},"N":"stop_gradient"}]},{"#":"1.dropout","A":[{"AT":{"#":"0.a_bool","D":true},"N":"is_test"},{"AT":{"#":"0.a_str","D":"upscale_in_train"},"N":"mode"},{"AT":{"#":"0.a_i32","D":0},"N":"seed"},{"AT":{"#":"0.a_bool","D":false},"N":"fix_seed"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_5/Dropout/"},"N":"struct_name"}],"I":[{"%":505},{"%":0},{"%":506}],"O":[{"%":507,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512,768],"NCHW",[],0]}},{"%":508,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_ui8"},[-1,512,768],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false},{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.add","A":[{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_5/"},"N":"struct_name"}],"I":[{"%":474},{"%":507}],"O":[{"%":509,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512,768],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.layer_norm","A":[{"AT":{"#":"0.a_f32","D":9.999999960041972e-13},"N":"epsilon"},{"AT":{"#":"0.a_i32","D":2},"N":"begin_norm_axis"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_5/LayerNorm/"},"N":"struct_name"}],"I":[{"%":509},{"%":104},{"%":103}],"O":[{"%":510,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512,768],"NCHW",[],0]}},{"%":511,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512],"NCHW",[],0]}},{"%":512,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false},{"#":"0.a_bool","D":false},{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.matmul","A":[{"AT":{"#":"0.a_bool","D":false},"N":"transpose_x"},{"AT":{"#":"0.a_bool","D":false},"N":"transpose_y"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_5/Linear/"},"N":"struct_name"}],"I":[{"%":510},{"%":108}],"O":[{"%":513,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512,3072],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.add","A":[{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_5/Linear/"},"N":"struct_name"}],"I":[{"%":513},{"%":107}],"O":[{"%":514,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512,3072],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.gelu","A":[{"AT":{"#":"0.a_bool","D":false},"N":"approximate"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_5/"},"N":"struct_name"}],"I":[{"%":514}],"O":[{"%":515,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512,3072],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.matmul","A":[{"AT":{"#":"0.a_bool","D":false},"N":"transpose_x"},{"AT":{"#":"0.a_bool","D":false},"N":"transpose_y"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_5/Linear_1/"},"N":"struct_name"}],"I":[{"%":515},{"%":106}],"O":[{"%":516,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512,768],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.add","A":[{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_5/Linear_1/"},"N":"struct_name"}],"I":[{"%":516},{"%":105}],"O":[{"%":517,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512,768],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.full","A":[{"AT":{"#":"1.a_intarray","D":[1]},"N":"shape"},{"AT":{"#":"0.a_f64","D":0.10000000149011612},"N":"value"},{"AT":{"#":"1.a_dtype","D":"float32"},"N":"dtype"},{"AT":{"#":"1.a_place","D":[1,0,""]},"N":"place"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_5/Dropout_2/"},"N":"struct_name"}],"I":[],"O":[{"%":518,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[1],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":true}]},"N":"stop_gradient"}]},{"#":"1.dropout","A":[{"AT":{"#":"0.a_bool","D":true},"N":"is_test"},{"AT":{"#":"0.a_str","D":"upscale_in_train"},"N":"mode"},{"AT":{"#":"0.a_i32","D":0},"N":"seed"},{"AT":{"#":"0.a_bool","D":false},"N":"fix_seed"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_5/Dropout_2/"},"N":"struct_name"}],"I":[{"%":517},{"%":0},{"%":518}],"O":[{"%":519,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512,768],"NCHW",[],0]}},{"%":520,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_ui8"},[-1,512,768],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false},{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.add","A":[{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_5/"},"N":"struct_name"}],"I":[{"%":510},{"%":519}],"O":[{"%":521,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512,768],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.layer_norm","A":[{"AT":{"#":"0.a_f32","D":9.999999960041972e-13},"N":"epsilon"},{"AT":{"#":"0.a_i32","D":2},"N":"begin_norm_axis"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_5/LayerNorm_1/"},"N":"struct_name"}],"I":[{"%":521},{"%":102},{"%":101}],"O":[{"%":522,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512,768],"NCHW",[],0]}},{"%":523,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512],"NCHW",[],0]}},{"%":524,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false},{"#":"0.a_bool","D":false},{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.matmul","A":[{"AT":{"#":"0.a_bool","D":false},"N":"transpose_x"},{"AT":{"#":"0.a_bool","D":false},"N":"transpose_y"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_6/MultiHeadAttention/Linear/"},"N":"struct_name"}],"I":[{"%":522},{"%":100}],"O":[{"%":525,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512,768],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.add","A":[{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_6/MultiHeadAttention/Linear/"},"N":"struct_name"}],"I":[{"%":525},{"%":99}],"O":[{"%":526,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512,768],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.full_int_array","A":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_i64","D":0},{"#":"0.a_i64","D":0},{"#":"0.a_i64","D":12},{"#":"0.a_i64","D":64}]},"N":"value"},{"AT":{"#":"1.a_dtype","D":"int64"},"N":"dtype"},{"AT":{"#":"1.a_place","D":[1,0,""]},"N":"place"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_6/MultiHeadAttention/"},"N":"struct_name"}],"I":[],"O":[{"%":527,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_i64"},[4],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":true}]},"N":"stop_gradient"}]},{"#":"1.reshape","A":[{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_6/MultiHeadAttention/"},"N":"struct_name"}],"I":[{"%":526},{"%":527}],"O":[{"%":528,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512,12,64],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.transpose","A":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_i32","D":0},{"#":"0.a_i32","D":2},{"#":"0.a_i32","D":1},{"#":"0.a_i32","D":3}]},"N":"perm"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_6/MultiHeadAttention/"},"N":"struct_name"}],"I":[{"%":528}],"O":[{"%":529,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,12,512,64],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.matmul","A":[{"AT":{"#":"0.a_bool","D":false},"N":"transpose_x"},{"AT":{"#":"0.a_bool","D":false},"N":"transpose_y"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_6/MultiHeadAttention/Linear_1/"},"N":"struct_name"}],"I":[{"%":522},{"%":98}],"O":[{"%":530,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512,768],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.add","A":[{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_6/MultiHeadAttention/Linear_1/"},"N":"struct_name"}],"I":[{"%":530},{"%":97}],"O":[{"%":531,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512,768],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.matmul","A":[{"AT":{"#":"0.a_bool","D":false},"N":"transpose_x"},{"AT":{"#":"0.a_bool","D":false},"N":"transpose_y"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_6/MultiHeadAttention/Linear_2/"},"N":"struct_name"}],"I":[{"%":522},{"%":96}],"O":[{"%":532,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512,768],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.add","A":[{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_6/MultiHeadAttention/Linear_2/"},"N":"struct_name"}],"I":[{"%":532},{"%":95}],"O":[{"%":533,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512,768],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.full_int_array","A":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_i64","D":0},{"#":"0.a_i64","D":0},{"#":"0.a_i64","D":12},{"#":"0.a_i64","D":64}]},"N":"value"},{"AT":{"#":"1.a_dtype","D":"int64"},"N":"dtype"},{"AT":{"#":"1.a_place","D":[1,0,""]},"N":"place"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_6/MultiHeadAttention/"},"N":"struct_name"}],"I":[],"O":[{"%":534,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_i64"},[4],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":true}]},"N":"stop_gradient"}]},{"#":"1.reshape","A":[{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_6/MultiHeadAttention/"},"N":"struct_name"}],"I":[{"%":531},{"%":534}],"O":[{"%":535,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512,12,64],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.transpose","A":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_i32","D":0},{"#":"0.a_i32","D":2},{"#":"0.a_i32","D":1},{"#":"0.a_i32","D":3}]},"N":"perm"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_6/MultiHeadAttention/"},"N":"struct_name"}],"I":[{"%":535}],"O":[{"%":536,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,12,512,64],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.full_int_array","A":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_i64","D":0},{"#":"0.a_i64","D":0},{"#":"0.a_i64","D":12},{"#":"0.a_i64","D":64}]},"N":"value"},{"AT":{"#":"1.a_dtype","D":"int64"},"N":"dtype"},{"AT":{"#":"1.a_place","D":[1,0,""]},"N":"place"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_6/MultiHeadAttention/"},"N":"struct_name"}],"I":[],"O":[{"%":537,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_i64"},[4],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":true}]},"N":"stop_gradient"}]},{"#":"1.reshape","A":[{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_6/MultiHeadAttention/"},"N":"struct_name"}],"I":[{"%":533},{"%":537}],"O":[{"%":538,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512,12,64],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.transpose","A":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_i32","D":0},{"#":"0.a_i32","D":2},{"#":"0.a_i32","D":1},{"#":"0.a_i32","D":3}]},"N":"perm"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_6/MultiHeadAttention/"},"N":"struct_name"}],"I":[{"%":538}],"O":[{"%":539,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,12,512,64],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.full","A":[{"AT":{"#":"1.a_intarray","D":[1]},"N":"shape"},{"AT":{"#":"0.a_f64","D":0.125},"N":"value"},{"AT":{"#":"1.a_dtype","D":"float32"},"N":"dtype"},{"AT":{"#":"1.a_place","D":[1,0,""]},"N":"place"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_6/MultiHeadAttention/"},"N":"struct_name"}],"I":[],"O":[{"%":540,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[1],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":true}]},"N":"stop_gradient"}]},{"#":"1.scale","A":[{"AT":{"#":"0.a_f32","D":0.0},"N":"bias"},{"AT":{"#":"0.a_bool","D":true},"N":"bias_after_scale"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_6/MultiHeadAttention/"},"N":"struct_name"}],"I":[{"%":529},{"%":540}],"O":[{"%":541,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,12,512,64],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.matmul","A":[{"AT":{"#":"0.a_bool","D":false},"N":"transpose_x"},{"AT":{"#":"0.a_bool","D":true},"N":"transpose_y"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_6/MultiHeadAttention/"},"N":"struct_name"}],"I":[{"%":541},{"%":536}],"O":[{"%":542,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,12,512,512],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.add","A":[{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_6/MultiHeadAttention/"},"N":"struct_name"}],"I":[{"%":542},{"%":212}],"O":[{"%":543,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,12,512,512],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.softmax","A":[{"AT":{"#":"0.a_i32","D":-1},"N":"axis"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_6/MultiHeadAttention/"},"N":"struct_name"}],"I":[{"%":543}],"O":[{"%":544,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,12,512,512],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.full","A":[{"AT":{"#":"1.a_intarray","D":[1]},"N":"shape"},{"AT":{"#":"0.a_f64","D":0.10000000149011612},"N":"value"},{"AT":{"#":"1.a_dtype","D":"float32"},"N":"dtype"},{"AT":{"#":"1.a_place","D":[1,0,""]},"N":"place"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_6/MultiHeadAttention/"},"N":"struct_name"}],"I":[],"O":[{"%":545,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[1],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":true}]},"N":"stop_gradient"}]},{"#":"1.dropout","A":[{"AT":{"#":"0.a_bool","D":true},"N":"is_test"},{"AT":{"#":"0.a_str","D":"upscale_in_train"},"N":"mode"},{"AT":{"#":"0.a_i32","D":0},"N":"seed"},{"AT":{"#":"0.a_bool","D":false},"N":"fix_seed"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_6/MultiHeadAttention/"},"N":"struct_name"}],"I":[{"%":544},{"%":0},{"%":545}],"O":[{"%":546,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,12,512,512],"NCHW",[],0]}},{"%":547,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_ui8"},[-1,12,512,512],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false},{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.matmul","A":[{"AT":{"#":"0.a_bool","D":false},"N":"transpose_x"},{"AT":{"#":"0.a_bool","D":false},"N":"transpose_y"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_6/MultiHeadAttention/"},"N":"struct_name"}],"I":[{"%":546},{"%":539}],"O":[{"%":548,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,12,512,64],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.transpose","A":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_i32","D":0},{"#":"0.a_i32","D":2},{"#":"0.a_i32","D":1},{"#":"0.a_i32","D":3}]},"N":"perm"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_6/MultiHeadAttention/"},"N":"struct_name"}],"I":[{"%":548}],"O":[{"%":549,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512,12,64],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.full_int_array","A":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_i64","D":0},{"#":"0.a_i64","D":0},{"#":"0.a_i64","D":768}]},"N":"value"},{"AT":{"#":"1.a_dtype","D":"int64"},"N":"dtype"},{"AT":{"#":"1.a_place","D":[1,0,""]},"N":"place"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_6/MultiHeadAttention/"},"N":"struct_name"}],"I":[],"O":[{"%":550,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_i64"},[3],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":true}]},"N":"stop_gradient"}]},{"#":"1.reshape","A":[{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_6/MultiHeadAttention/"},"N":"struct_name"}],"I":[{"%":549},{"%":550}],"O":[{"%":551,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512,768],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.matmul","A":[{"AT":{"#":"0.a_bool","D":false},"N":"transpose_x"},{"AT":{"#":"0.a_bool","D":false},"N":"transpose_y"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_6/MultiHeadAttention/Linear_3/"},"N":"struct_name"}],"I":[{"%":551},{"%":94}],"O":[{"%":552,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512,768],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.add","A":[{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_6/MultiHeadAttention/Linear_3/"},"N":"struct_name"}],"I":[{"%":552},{"%":93}],"O":[{"%":553,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512,768],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.full","A":[{"AT":{"#":"1.a_intarray","D":[1]},"N":"shape"},{"AT":{"#":"0.a_f64","D":0.10000000149011612},"N":"value"},{"AT":{"#":"1.a_dtype","D":"float32"},"N":"dtype"},{"AT":{"#":"1.a_place","D":[1,0,""]},"N":"place"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_6/Dropout/"},"N":"struct_name"}],"I":[],"O":[{"%":554,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[1],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":true}]},"N":"stop_gradient"}]},{"#":"1.dropout","A":[{"AT":{"#":"0.a_bool","D":true},"N":"is_test"},{"AT":{"#":"0.a_str","D":"upscale_in_train"},"N":"mode"},{"AT":{"#":"0.a_i32","D":0},"N":"seed"},{"AT":{"#":"0.a_bool","D":false},"N":"fix_seed"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_6/Dropout/"},"N":"struct_name"}],"I":[{"%":553},{"%":0},{"%":554}],"O":[{"%":555,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512,768],"NCHW",[],0]}},{"%":556,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_ui8"},[-1,512,768],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false},{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.add","A":[{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_6/"},"N":"struct_name"}],"I":[{"%":522},{"%":555}],"O":[{"%":557,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512,768],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.layer_norm","A":[{"AT":{"#":"0.a_f32","D":9.999999960041972e-13},"N":"epsilon"},{"AT":{"#":"0.a_i32","D":2},"N":"begin_norm_axis"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_6/LayerNorm/"},"N":"struct_name"}],"I":[{"%":557},{"%":88},{"%":87}],"O":[{"%":558,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512,768],"NCHW",[],0]}},{"%":559,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512],"NCHW",[],0]}},{"%":560,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false},{"#":"0.a_bool","D":false},{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.matmul","A":[{"AT":{"#":"0.a_bool","D":false},"N":"transpose_x"},{"AT":{"#":"0.a_bool","D":false},"N":"transpose_y"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_6/Linear/"},"N":"struct_name"}],"I":[{"%":558},{"%":92}],"O":[{"%":561,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512,3072],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.add","A":[{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_6/Linear/"},"N":"struct_name"}],"I":[{"%":561},{"%":91}],"O":[{"%":562,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512,3072],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.gelu","A":[{"AT":{"#":"0.a_bool","D":false},"N":"approximate"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_6/"},"N":"struct_name"}],"I":[{"%":562}],"O":[{"%":563,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512,3072],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.matmul","A":[{"AT":{"#":"0.a_bool","D":false},"N":"transpose_x"},{"AT":{"#":"0.a_bool","D":false},"N":"transpose_y"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_6/Linear_1/"},"N":"struct_name"}],"I":[{"%":563},{"%":90}],"O":[{"%":564,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512,768],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.add","A":[{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_6/Linear_1/"},"N":"struct_name"}],"I":[{"%":564},{"%":89}],"O":[{"%":565,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512,768],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.full","A":[{"AT":{"#":"1.a_intarray","D":[1]},"N":"shape"},{"AT":{"#":"0.a_f64","D":0.10000000149011612},"N":"value"},{"AT":{"#":"1.a_dtype","D":"float32"},"N":"dtype"},{"AT":{"#":"1.a_place","D":[1,0,""]},"N":"place"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_6/Dropout_2/"},"N":"struct_name"}],"I":[],"O":[{"%":566,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[1],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":true}]},"N":"stop_gradient"}]},{"#":"1.dropout","A":[{"AT":{"#":"0.a_bool","D":true},"N":"is_test"},{"AT":{"#":"0.a_str","D":"upscale_in_train"},"N":"mode"},{"AT":{"#":"0.a_i32","D":0},"N":"seed"},{"AT":{"#":"0.a_bool","D":false},"N":"fix_seed"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_6/Dropout_2/"},"N":"struct_name"}],"I":[{"%":565},{"%":0},{"%":566}],"O":[{"%":567,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512,768],"NCHW",[],0]}},{"%":568,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_ui8"},[-1,512,768],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false},{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.add","A":[{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_6/"},"N":"struct_name"}],"I":[{"%":558},{"%":567}],"O":[{"%":569,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512,768],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.layer_norm","A":[{"AT":{"#":"0.a_f32","D":9.999999960041972e-13},"N":"epsilon"},{"AT":{"#":"0.a_i32","D":2},"N":"begin_norm_axis"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_6/LayerNorm_1/"},"N":"struct_name"}],"I":[{"%":569},{"%":86},{"%":85}],"O":[{"%":570,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512,768],"NCHW",[],0]}},{"%":571,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512],"NCHW",[],0]}},{"%":572,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false},{"#":"0.a_bool","D":false},{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.matmul","A":[{"AT":{"#":"0.a_bool","D":false},"N":"transpose_x"},{"AT":{"#":"0.a_bool","D":false},"N":"transpose_y"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_7/MultiHeadAttention/Linear/"},"N":"struct_name"}],"I":[{"%":570},{"%":84}],"O":[{"%":573,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512,768],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.add","A":[{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_7/MultiHeadAttention/Linear/"},"N":"struct_name"}],"I":[{"%":573},{"%":83}],"O":[{"%":574,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512,768],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.full_int_array","A":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_i64","D":0},{"#":"0.a_i64","D":0},{"#":"0.a_i64","D":12},{"#":"0.a_i64","D":64}]},"N":"value"},{"AT":{"#":"1.a_dtype","D":"int64"},"N":"dtype"},{"AT":{"#":"1.a_place","D":[1,0,""]},"N":"place"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_7/MultiHeadAttention/"},"N":"struct_name"}],"I":[],"O":[{"%":575,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_i64"},[4],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":true}]},"N":"stop_gradient"}]},{"#":"1.reshape","A":[{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_7/MultiHeadAttention/"},"N":"struct_name"}],"I":[{"%":574},{"%":575}],"O":[{"%":576,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512,12,64],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.transpose","A":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_i32","D":0},{"#":"0.a_i32","D":2},{"#":"0.a_i32","D":1},{"#":"0.a_i32","D":3}]},"N":"perm"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_7/MultiHeadAttention/"},"N":"struct_name"}],"I":[{"%":576}],"O":[{"%":577,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,12,512,64],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.matmul","A":[{"AT":{"#":"0.a_bool","D":false},"N":"transpose_x"},{"AT":{"#":"0.a_bool","D":false},"N":"transpose_y"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_7/MultiHeadAttention/Linear_1/"},"N":"struct_name"}],"I":[{"%":570},{"%":82}],"O":[{"%":578,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512,768],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.add","A":[{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_7/MultiHeadAttention/Linear_1/"},"N":"struct_name"}],"I":[{"%":578},{"%":81}],"O":[{"%":579,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512,768],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.matmul","A":[{"AT":{"#":"0.a_bool","D":false},"N":"transpose_x"},{"AT":{"#":"0.a_bool","D":false},"N":"transpose_y"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_7/MultiHeadAttention/Linear_2/"},"N":"struct_name"}],"I":[{"%":570},{"%":80}],"O":[{"%":580,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512,768],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.add","A":[{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_7/MultiHeadAttention/Linear_2/"},"N":"struct_name"}],"I":[{"%":580},{"%":79}],"O":[{"%":581,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512,768],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.full_int_array","A":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_i64","D":0},{"#":"0.a_i64","D":0},{"#":"0.a_i64","D":12},{"#":"0.a_i64","D":64}]},"N":"value"},{"AT":{"#":"1.a_dtype","D":"int64"},"N":"dtype"},{"AT":{"#":"1.a_place","D":[1,0,""]},"N":"place"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_7/MultiHeadAttention/"},"N":"struct_name"}],"I":[],"O":[{"%":582,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_i64"},[4],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":true}]},"N":"stop_gradient"}]},{"#":"1.reshape","A":[{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_7/MultiHeadAttention/"},"N":"struct_name"}],"I":[{"%":579},{"%":582}],"O":[{"%":583,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512,12,64],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.transpose","A":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_i32","D":0},{"#":"0.a_i32","D":2},{"#":"0.a_i32","D":1},{"#":"0.a_i32","D":3}]},"N":"perm"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_7/MultiHeadAttention/"},"N":"struct_name"}],"I":[{"%":583}],"O":[{"%":584,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,12,512,64],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.full_int_array","A":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_i64","D":0},{"#":"0.a_i64","D":0},{"#":"0.a_i64","D":12},{"#":"0.a_i64","D":64}]},"N":"value"},{"AT":{"#":"1.a_dtype","D":"int64"},"N":"dtype"},{"AT":{"#":"1.a_place","D":[1,0,""]},"N":"place"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_7/MultiHeadAttention/"},"N":"struct_name"}],"I":[],"O":[{"%":585,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_i64"},[4],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":true}]},"N":"stop_gradient"}]},{"#":"1.reshape","A":[{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_7/MultiHeadAttention/"},"N":"struct_name"}],"I":[{"%":581},{"%":585}],"O":[{"%":586,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512,12,64],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.transpose","A":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_i32","D":0},{"#":"0.a_i32","D":2},{"#":"0.a_i32","D":1},{"#":"0.a_i32","D":3}]},"N":"perm"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_7/MultiHeadAttention/"},"N":"struct_name"}],"I":[{"%":586}],"O":[{"%":587,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,12,512,64],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.full","A":[{"AT":{"#":"1.a_intarray","D":[1]},"N":"shape"},{"AT":{"#":"0.a_f64","D":0.125},"N":"value"},{"AT":{"#":"1.a_dtype","D":"float32"},"N":"dtype"},{"AT":{"#":"1.a_place","D":[1,0,""]},"N":"place"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_7/MultiHeadAttention/"},"N":"struct_name"}],"I":[],"O":[{"%":588,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[1],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":true}]},"N":"stop_gradient"}]},{"#":"1.scale","A":[{"AT":{"#":"0.a_f32","D":0.0},"N":"bias"},{"AT":{"#":"0.a_bool","D":true},"N":"bias_after_scale"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_7/MultiHeadAttention/"},"N":"struct_name"}],"I":[{"%":577},{"%":588}],"O":[{"%":589,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,12,512,64],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.matmul","A":[{"AT":{"#":"0.a_bool","D":false},"N":"transpose_x"},{"AT":{"#":"0.a_bool","D":true},"N":"transpose_y"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_7/MultiHeadAttention/"},"N":"struct_name"}],"I":[{"%":589},{"%":584}],"O":[{"%":590,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,12,512,512],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.add","A":[{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_7/MultiHeadAttention/"},"N":"struct_name"}],"I":[{"%":590},{"%":212}],"O":[{"%":591,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,12,512,512],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.softmax","A":[{"AT":{"#":"0.a_i32","D":-1},"N":"axis"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_7/MultiHeadAttention/"},"N":"struct_name"}],"I":[{"%":591}],"O":[{"%":592,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,12,512,512],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.full","A":[{"AT":{"#":"1.a_intarray","D":[1]},"N":"shape"},{"AT":{"#":"0.a_f64","D":0.10000000149011612},"N":"value"},{"AT":{"#":"1.a_dtype","D":"float32"},"N":"dtype"},{"AT":{"#":"1.a_place","D":[1,0,""]},"N":"place"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_7/MultiHeadAttention/"},"N":"struct_name"}],"I":[],"O":[{"%":593,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[1],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":true}]},"N":"stop_gradient"}]},{"#":"1.dropout","A":[{"AT":{"#":"0.a_bool","D":true},"N":"is_test"},{"AT":{"#":"0.a_str","D":"upscale_in_train"},"N":"mode"},{"AT":{"#":"0.a_i32","D":0},"N":"seed"},{"AT":{"#":"0.a_bool","D":false},"N":"fix_seed"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_7/MultiHeadAttention/"},"N":"struct_name"}],"I":[{"%":592},{"%":0},{"%":593}],"O":[{"%":594,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,12,512,512],"NCHW",[],0]}},{"%":595,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_ui8"},[-1,12,512,512],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false},{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.matmul","A":[{"AT":{"#":"0.a_bool","D":false},"N":"transpose_x"},{"AT":{"#":"0.a_bool","D":false},"N":"transpose_y"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_7/MultiHeadAttention/"},"N":"struct_name"}],"I":[{"%":594},{"%":587}],"O":[{"%":596,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,12,512,64],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.transpose","A":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_i32","D":0},{"#":"0.a_i32","D":2},{"#":"0.a_i32","D":1},{"#":"0.a_i32","D":3}]},"N":"perm"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_7/MultiHeadAttention/"},"N":"struct_name"}],"I":[{"%":596}],"O":[{"%":597,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512,12,64],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.full_int_array","A":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_i64","D":0},{"#":"0.a_i64","D":0},{"#":"0.a_i64","D":768}]},"N":"value"},{"AT":{"#":"1.a_dtype","D":"int64"},"N":"dtype"},{"AT":{"#":"1.a_place","D":[1,0,""]},"N":"place"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_7/MultiHeadAttention/"},"N":"struct_name"}],"I":[],"O":[{"%":598,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_i64"},[3],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":true}]},"N":"stop_gradient"}]},{"#":"1.reshape","A":[{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_7/MultiHeadAttention/"},"N":"struct_name"}],"I":[{"%":597},{"%":598}],"O":[{"%":599,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512,768],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.matmul","A":[{"AT":{"#":"0.a_bool","D":false},"N":"transpose_x"},{"AT":{"#":"0.a_bool","D":false},"N":"transpose_y"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_7/MultiHeadAttention/Linear_3/"},"N":"struct_name"}],"I":[{"%":599},{"%":78}],"O":[{"%":600,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512,768],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.add","A":[{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_7/MultiHeadAttention/Linear_3/"},"N":"struct_name"}],"I":[{"%":600},{"%":77}],"O":[{"%":601,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512,768],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.full","A":[{"AT":{"#":"1.a_intarray","D":[1]},"N":"shape"},{"AT":{"#":"0.a_f64","D":0.10000000149011612},"N":"value"},{"AT":{"#":"1.a_dtype","D":"float32"},"N":"dtype"},{"AT":{"#":"1.a_place","D":[1,0,""]},"N":"place"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_7/Dropout/"},"N":"struct_name"}],"I":[],"O":[{"%":602,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[1],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":true}]},"N":"stop_gradient"}]},{"#":"1.dropout","A":[{"AT":{"#":"0.a_bool","D":true},"N":"is_test"},{"AT":{"#":"0.a_str","D":"upscale_in_train"},"N":"mode"},{"AT":{"#":"0.a_i32","D":0},"N":"seed"},{"AT":{"#":"0.a_bool","D":false},"N":"fix_seed"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_7/Dropout/"},"N":"struct_name"}],"I":[{"%":601},{"%":0},{"%":602}],"O":[{"%":603,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512,768],"NCHW",[],0]}},{"%":604,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_ui8"},[-1,512,768],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false},{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.add","A":[{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_7/"},"N":"struct_name"}],"I":[{"%":570},{"%":603}],"O":[{"%":605,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512,768],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.layer_norm","A":[{"AT":{"#":"0.a_f32","D":9.999999960041972e-13},"N":"epsilon"},{"AT":{"#":"0.a_i32","D":2},"N":"begin_norm_axis"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_7/LayerNorm/"},"N":"struct_name"}],"I":[{"%":605},{"%":72},{"%":71}],"O":[{"%":606,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512,768],"NCHW",[],0]}},{"%":607,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512],"NCHW",[],0]}},{"%":608,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false},{"#":"0.a_bool","D":false},{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.matmul","A":[{"AT":{"#":"0.a_bool","D":false},"N":"transpose_x"},{"AT":{"#":"0.a_bool","D":false},"N":"transpose_y"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_7/Linear/"},"N":"struct_name"}],"I":[{"%":606},{"%":76}],"O":[{"%":609,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512,3072],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.add","A":[{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_7/Linear/"},"N":"struct_name"}],"I":[{"%":609},{"%":75}],"O":[{"%":610,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512,3072],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.gelu","A":[{"AT":{"#":"0.a_bool","D":false},"N":"approximate"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_7/"},"N":"struct_name"}],"I":[{"%":610}],"O":[{"%":611,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512,3072],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.matmul","A":[{"AT":{"#":"0.a_bool","D":false},"N":"transpose_x"},{"AT":{"#":"0.a_bool","D":false},"N":"transpose_y"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_7/Linear_1/"},"N":"struct_name"}],"I":[{"%":611},{"%":74}],"O":[{"%":612,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512,768],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.add","A":[{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_7/Linear_1/"},"N":"struct_name"}],"I":[{"%":612},{"%":73}],"O":[{"%":613,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512,768],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.full","A":[{"AT":{"#":"1.a_intarray","D":[1]},"N":"shape"},{"AT":{"#":"0.a_f64","D":0.10000000149011612},"N":"value"},{"AT":{"#":"1.a_dtype","D":"float32"},"N":"dtype"},{"AT":{"#":"1.a_place","D":[1,0,""]},"N":"place"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_7/Dropout_2/"},"N":"struct_name"}],"I":[],"O":[{"%":614,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[1],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":true}]},"N":"stop_gradient"}]},{"#":"1.dropout","A":[{"AT":{"#":"0.a_bool","D":true},"N":"is_test"},{"AT":{"#":"0.a_str","D":"upscale_in_train"},"N":"mode"},{"AT":{"#":"0.a_i32","D":0},"N":"seed"},{"AT":{"#":"0.a_bool","D":false},"N":"fix_seed"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_7/Dropout_2/"},"N":"struct_name"}],"I":[{"%":613},{"%":0},{"%":614}],"O":[{"%":615,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512,768],"NCHW",[],0]}},{"%":616,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_ui8"},[-1,512,768],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false},{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.add","A":[{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_7/"},"N":"struct_name"}],"I":[{"%":606},{"%":615}],"O":[{"%":617,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512,768],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.layer_norm","A":[{"AT":{"#":"0.a_f32","D":9.999999960041972e-13},"N":"epsilon"},{"AT":{"#":"0.a_i32","D":2},"N":"begin_norm_axis"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_7/LayerNorm_1/"},"N":"struct_name"}],"I":[{"%":617},{"%":70},{"%":69}],"O":[{"%":618,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512,768],"NCHW",[],0]}},{"%":619,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512],"NCHW",[],0]}},{"%":620,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false},{"#":"0.a_bool","D":false},{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.matmul","A":[{"AT":{"#":"0.a_bool","D":false},"N":"transpose_x"},{"AT":{"#":"0.a_bool","D":false},"N":"transpose_y"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_8/MultiHeadAttention/Linear/"},"N":"struct_name"}],"I":[{"%":618},{"%":68}],"O":[{"%":621,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512,768],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.add","A":[{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_8/MultiHeadAttention/Linear/"},"N":"struct_name"}],"I":[{"%":621},{"%":67}],"O":[{"%":622,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512,768],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.full_int_array","A":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_i64","D":0},{"#":"0.a_i64","D":0},{"#":"0.a_i64","D":12},{"#":"0.a_i64","D":64}]},"N":"value"},{"AT":{"#":"1.a_dtype","D":"int64"},"N":"dtype"},{"AT":{"#":"1.a_place","D":[1,0,""]},"N":"place"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_8/MultiHeadAttention/"},"N":"struct_name"}],"I":[],"O":[{"%":623,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_i64"},[4],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":true}]},"N":"stop_gradient"}]},{"#":"1.reshape","A":[{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_8/MultiHeadAttention/"},"N":"struct_name"}],"I":[{"%":622},{"%":623}],"O":[{"%":624,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512,12,64],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.transpose","A":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_i32","D":0},{"#":"0.a_i32","D":2},{"#":"0.a_i32","D":1},{"#":"0.a_i32","D":3}]},"N":"perm"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_8/MultiHeadAttention/"},"N":"struct_name"}],"I":[{"%":624}],"O":[{"%":625,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,12,512,64],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.matmul","A":[{"AT":{"#":"0.a_bool","D":false},"N":"transpose_x"},{"AT":{"#":"0.a_bool","D":false},"N":"transpose_y"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_8/MultiHeadAttention/Linear_1/"},"N":"struct_name"}],"I":[{"%":618},{"%":66}],"O":[{"%":626,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512,768],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.add","A":[{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_8/MultiHeadAttention/Linear_1/"},"N":"struct_name"}],"I":[{"%":626},{"%":65}],"O":[{"%":627,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512,768],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.matmul","A":[{"AT":{"#":"0.a_bool","D":false},"N":"transpose_x"},{"AT":{"#":"0.a_bool","D":false},"N":"transpose_y"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_8/MultiHeadAttention/Linear_2/"},"N":"struct_name"}],"I":[{"%":618},{"%":64}],"O":[{"%":628,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512,768],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.add","A":[{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_8/MultiHeadAttention/Linear_2/"},"N":"struct_name"}],"I":[{"%":628},{"%":63}],"O":[{"%":629,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512,768],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.full_int_array","A":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_i64","D":0},{"#":"0.a_i64","D":0},{"#":"0.a_i64","D":12},{"#":"0.a_i64","D":64}]},"N":"value"},{"AT":{"#":"1.a_dtype","D":"int64"},"N":"dtype"},{"AT":{"#":"1.a_place","D":[1,0,""]},"N":"place"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_8/MultiHeadAttention/"},"N":"struct_name"}],"I":[],"O":[{"%":630,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_i64"},[4],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":true}]},"N":"stop_gradient"}]},{"#":"1.reshape","A":[{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_8/MultiHeadAttention/"},"N":"struct_name"}],"I":[{"%":627},{"%":630}],"O":[{"%":631,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512,12,64],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.transpose","A":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_i32","D":0},{"#":"0.a_i32","D":2},{"#":"0.a_i32","D":1},{"#":"0.a_i32","D":3}]},"N":"perm"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_8/MultiHeadAttention/"},"N":"struct_name"}],"I":[{"%":631}],"O":[{"%":632,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,12,512,64],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.full_int_array","A":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_i64","D":0},{"#":"0.a_i64","D":0},{"#":"0.a_i64","D":12},{"#":"0.a_i64","D":64}]},"N":"value"},{"AT":{"#":"1.a_dtype","D":"int64"},"N":"dtype"},{"AT":{"#":"1.a_place","D":[1,0,""]},"N":"place"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_8/MultiHeadAttention/"},"N":"struct_name"}],"I":[],"O":[{"%":633,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_i64"},[4],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":true}]},"N":"stop_gradient"}]},{"#":"1.reshape","A":[{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_8/MultiHeadAttention/"},"N":"struct_name"}],"I":[{"%":629},{"%":633}],"O":[{"%":634,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512,12,64],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.transpose","A":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_i32","D":0},{"#":"0.a_i32","D":2},{"#":"0.a_i32","D":1},{"#":"0.a_i32","D":3}]},"N":"perm"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_8/MultiHeadAttention/"},"N":"struct_name"}],"I":[{"%":634}],"O":[{"%":635,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,12,512,64],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.full","A":[{"AT":{"#":"1.a_intarray","D":[1]},"N":"shape"},{"AT":{"#":"0.a_f64","D":0.125},"N":"value"},{"AT":{"#":"1.a_dtype","D":"float32"},"N":"dtype"},{"AT":{"#":"1.a_place","D":[1,0,""]},"N":"place"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_8/MultiHeadAttention/"},"N":"struct_name"}],"I":[],"O":[{"%":636,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[1],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":true}]},"N":"stop_gradient"}]},{"#":"1.scale","A":[{"AT":{"#":"0.a_f32","D":0.0},"N":"bias"},{"AT":{"#":"0.a_bool","D":true},"N":"bias_after_scale"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_8/MultiHeadAttention/"},"N":"struct_name"}],"I":[{"%":625},{"%":636}],"O":[{"%":637,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,12,512,64],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.matmul","A":[{"AT":{"#":"0.a_bool","D":false},"N":"transpose_x"},{"AT":{"#":"0.a_bool","D":true},"N":"transpose_y"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_8/MultiHeadAttention/"},"N":"struct_name"}],"I":[{"%":637},{"%":632}],"O":[{"%":638,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,12,512,512],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.add","A":[{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_8/MultiHeadAttention/"},"N":"struct_name"}],"I":[{"%":638},{"%":212}],"O":[{"%":639,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,12,512,512],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.softmax","A":[{"AT":{"#":"0.a_i32","D":-1},"N":"axis"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_8/MultiHeadAttention/"},"N":"struct_name"}],"I":[{"%":639}],"O":[{"%":640,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,12,512,512],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.full","A":[{"AT":{"#":"1.a_intarray","D":[1]},"N":"shape"},{"AT":{"#":"0.a_f64","D":0.10000000149011612},"N":"value"},{"AT":{"#":"1.a_dtype","D":"float32"},"N":"dtype"},{"AT":{"#":"1.a_place","D":[1,0,""]},"N":"place"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_8/MultiHeadAttention/"},"N":"struct_name"}],"I":[],"O":[{"%":641,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[1],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":true}]},"N":"stop_gradient"}]},{"#":"1.dropout","A":[{"AT":{"#":"0.a_bool","D":true},"N":"is_test"},{"AT":{"#":"0.a_str","D":"upscale_in_train"},"N":"mode"},{"AT":{"#":"0.a_i32","D":0},"N":"seed"},{"AT":{"#":"0.a_bool","D":false},"N":"fix_seed"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_8/MultiHeadAttention/"},"N":"struct_name"}],"I":[{"%":640},{"%":0},{"%":641}],"O":[{"%":642,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,12,512,512],"NCHW",[],0]}},{"%":643,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_ui8"},[-1,12,512,512],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false},{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.matmul","A":[{"AT":{"#":"0.a_bool","D":false},"N":"transpose_x"},{"AT":{"#":"0.a_bool","D":false},"N":"transpose_y"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_8/MultiHeadAttention/"},"N":"struct_name"}],"I":[{"%":642},{"%":635}],"O":[{"%":644,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,12,512,64],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.transpose","A":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_i32","D":0},{"#":"0.a_i32","D":2},{"#":"0.a_i32","D":1},{"#":"0.a_i32","D":3}]},"N":"perm"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_8/MultiHeadAttention/"},"N":"struct_name"}],"I":[{"%":644}],"O":[{"%":645,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512,12,64],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.full_int_array","A":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_i64","D":0},{"#":"0.a_i64","D":0},{"#":"0.a_i64","D":768}]},"N":"value"},{"AT":{"#":"1.a_dtype","D":"int64"},"N":"dtype"},{"AT":{"#":"1.a_place","D":[1,0,""]},"N":"place"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_8/MultiHeadAttention/"},"N":"struct_name"}],"I":[],"O":[{"%":646,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_i64"},[3],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":true}]},"N":"stop_gradient"}]},{"#":"1.reshape","A":[{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_8/MultiHeadAttention/"},"N":"struct_name"}],"I":[{"%":645},{"%":646}],"O":[{"%":647,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512,768],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.matmul","A":[{"AT":{"#":"0.a_bool","D":false},"N":"transpose_x"},{"AT":{"#":"0.a_bool","D":false},"N":"transpose_y"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_8/MultiHeadAttention/Linear_3/"},"N":"struct_name"}],"I":[{"%":647},{"%":62}],"O":[{"%":648,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512,768],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.add","A":[{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_8/MultiHeadAttention/Linear_3/"},"N":"struct_name"}],"I":[{"%":648},{"%":61}],"O":[{"%":649,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512,768],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.full","A":[{"AT":{"#":"1.a_intarray","D":[1]},"N":"shape"},{"AT":{"#":"0.a_f64","D":0.10000000149011612},"N":"value"},{"AT":{"#":"1.a_dtype","D":"float32"},"N":"dtype"},{"AT":{"#":"1.a_place","D":[1,0,""]},"N":"place"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_8/Dropout/"},"N":"struct_name"}],"I":[],"O":[{"%":650,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[1],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":true}]},"N":"stop_gradient"}]},{"#":"1.dropout","A":[{"AT":{"#":"0.a_bool","D":true},"N":"is_test"},{"AT":{"#":"0.a_str","D":"upscale_in_train"},"N":"mode"},{"AT":{"#":"0.a_i32","D":0},"N":"seed"},{"AT":{"#":"0.a_bool","D":false},"N":"fix_seed"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_8/Dropout/"},"N":"struct_name"}],"I":[{"%":649},{"%":0},{"%":650}],"O":[{"%":651,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512,768],"NCHW",[],0]}},{"%":652,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_ui8"},[-1,512,768],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false},{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.add","A":[{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_8/"},"N":"struct_name"}],"I":[{"%":618},{"%":651}],"O":[{"%":653,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512,768],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.layer_norm","A":[{"AT":{"#":"0.a_f32","D":9.999999960041972e-13},"N":"epsilon"},{"AT":{"#":"0.a_i32","D":2},"N":"begin_norm_axis"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_8/LayerNorm/"},"N":"struct_name"}],"I":[{"%":653},{"%":56},{"%":55}],"O":[{"%":654,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512,768],"NCHW",[],0]}},{"%":655,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512],"NCHW",[],0]}},{"%":656,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false},{"#":"0.a_bool","D":false},{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.matmul","A":[{"AT":{"#":"0.a_bool","D":false},"N":"transpose_x"},{"AT":{"#":"0.a_bool","D":false},"N":"transpose_y"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_8/Linear/"},"N":"struct_name"}],"I":[{"%":654},{"%":60}],"O":[{"%":657,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512,3072],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.add","A":[{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_8/Linear/"},"N":"struct_name"}],"I":[{"%":657},{"%":59}],"O":[{"%":658,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512,3072],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.gelu","A":[{"AT":{"#":"0.a_bool","D":false},"N":"approximate"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_8/"},"N":"struct_name"}],"I":[{"%":658}],"O":[{"%":659,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512,3072],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.matmul","A":[{"AT":{"#":"0.a_bool","D":false},"N":"transpose_x"},{"AT":{"#":"0.a_bool","D":false},"N":"transpose_y"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_8/Linear_1/"},"N":"struct_name"}],"I":[{"%":659},{"%":58}],"O":[{"%":660,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512,768],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.add","A":[{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_8/Linear_1/"},"N":"struct_name"}],"I":[{"%":660},{"%":57}],"O":[{"%":661,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512,768],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.full","A":[{"AT":{"#":"1.a_intarray","D":[1]},"N":"shape"},{"AT":{"#":"0.a_f64","D":0.10000000149011612},"N":"value"},{"AT":{"#":"1.a_dtype","D":"float32"},"N":"dtype"},{"AT":{"#":"1.a_place","D":[1,0,""]},"N":"place"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_8/Dropout_2/"},"N":"struct_name"}],"I":[],"O":[{"%":662,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[1],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":true}]},"N":"stop_gradient"}]},{"#":"1.dropout","A":[{"AT":{"#":"0.a_bool","D":true},"N":"is_test"},{"AT":{"#":"0.a_str","D":"upscale_in_train"},"N":"mode"},{"AT":{"#":"0.a_i32","D":0},"N":"seed"},{"AT":{"#":"0.a_bool","D":false},"N":"fix_seed"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_8/Dropout_2/"},"N":"struct_name"}],"I":[{"%":661},{"%":0},{"%":662}],"O":[{"%":663,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512,768],"NCHW",[],0]}},{"%":664,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_ui8"},[-1,512,768],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false},{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.add","A":[{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_8/"},"N":"struct_name"}],"I":[{"%":654},{"%":663}],"O":[{"%":665,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512,768],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.layer_norm","A":[{"AT":{"#":"0.a_f32","D":9.999999960041972e-13},"N":"epsilon"},{"AT":{"#":"0.a_i32","D":2},"N":"begin_norm_axis"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_8/LayerNorm_1/"},"N":"struct_name"}],"I":[{"%":665},{"%":54},{"%":53}],"O":[{"%":666,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512,768],"NCHW",[],0]}},{"%":667,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512],"NCHW",[],0]}},{"%":668,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false},{"#":"0.a_bool","D":false},{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.matmul","A":[{"AT":{"#":"0.a_bool","D":false},"N":"transpose_x"},{"AT":{"#":"0.a_bool","D":false},"N":"transpose_y"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_9/MultiHeadAttention/Linear/"},"N":"struct_name"}],"I":[{"%":666},{"%":52}],"O":[{"%":669,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512,768],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.add","A":[{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_9/MultiHeadAttention/Linear/"},"N":"struct_name"}],"I":[{"%":669},{"%":51}],"O":[{"%":670,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512,768],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.full_int_array","A":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_i64","D":0},{"#":"0.a_i64","D":0},{"#":"0.a_i64","D":12},{"#":"0.a_i64","D":64}]},"N":"value"},{"AT":{"#":"1.a_dtype","D":"int64"},"N":"dtype"},{"AT":{"#":"1.a_place","D":[1,0,""]},"N":"place"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_9/MultiHeadAttention/"},"N":"struct_name"}],"I":[],"O":[{"%":671,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_i64"},[4],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":true}]},"N":"stop_gradient"}]},{"#":"1.reshape","A":[{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_9/MultiHeadAttention/"},"N":"struct_name"}],"I":[{"%":670},{"%":671}],"O":[{"%":672,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512,12,64],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.transpose","A":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_i32","D":0},{"#":"0.a_i32","D":2},{"#":"0.a_i32","D":1},{"#":"0.a_i32","D":3}]},"N":"perm"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_9/MultiHeadAttention/"},"N":"struct_name"}],"I":[{"%":672}],"O":[{"%":673,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,12,512,64],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.matmul","A":[{"AT":{"#":"0.a_bool","D":false},"N":"transpose_x"},{"AT":{"#":"0.a_bool","D":false},"N":"transpose_y"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_9/MultiHeadAttention/Linear_1/"},"N":"struct_name"}],"I":[{"%":666},{"%":50}],"O":[{"%":674,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512,768],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.add","A":[{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_9/MultiHeadAttention/Linear_1/"},"N":"struct_name"}],"I":[{"%":674},{"%":49}],"O":[{"%":675,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512,768],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.matmul","A":[{"AT":{"#":"0.a_bool","D":false},"N":"transpose_x"},{"AT":{"#":"0.a_bool","D":false},"N":"transpose_y"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_9/MultiHeadAttention/Linear_2/"},"N":"struct_name"}],"I":[{"%":666},{"%":48}],"O":[{"%":676,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512,768],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.add","A":[{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_9/MultiHeadAttention/Linear_2/"},"N":"struct_name"}],"I":[{"%":676},{"%":47}],"O":[{"%":677,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512,768],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.full_int_array","A":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_i64","D":0},{"#":"0.a_i64","D":0},{"#":"0.a_i64","D":12},{"#":"0.a_i64","D":64}]},"N":"value"},{"AT":{"#":"1.a_dtype","D":"int64"},"N":"dtype"},{"AT":{"#":"1.a_place","D":[1,0,""]},"N":"place"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_9/MultiHeadAttention/"},"N":"struct_name"}],"I":[],"O":[{"%":678,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_i64"},[4],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":true}]},"N":"stop_gradient"}]},{"#":"1.reshape","A":[{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_9/MultiHeadAttention/"},"N":"struct_name"}],"I":[{"%":675},{"%":678}],"O":[{"%":679,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512,12,64],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.transpose","A":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_i32","D":0},{"#":"0.a_i32","D":2},{"#":"0.a_i32","D":1},{"#":"0.a_i32","D":3}]},"N":"perm"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_9/MultiHeadAttention/"},"N":"struct_name"}],"I":[{"%":679}],"O":[{"%":680,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,12,512,64],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.full_int_array","A":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_i64","D":0},{"#":"0.a_i64","D":0},{"#":"0.a_i64","D":12},{"#":"0.a_i64","D":64}]},"N":"value"},{"AT":{"#":"1.a_dtype","D":"int64"},"N":"dtype"},{"AT":{"#":"1.a_place","D":[1,0,""]},"N":"place"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_9/MultiHeadAttention/"},"N":"struct_name"}],"I":[],"O":[{"%":681,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_i64"},[4],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":true}]},"N":"stop_gradient"}]},{"#":"1.reshape","A":[{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_9/MultiHeadAttention/"},"N":"struct_name"}],"I":[{"%":677},{"%":681}],"O":[{"%":682,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512,12,64],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.transpose","A":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_i32","D":0},{"#":"0.a_i32","D":2},{"#":"0.a_i32","D":1},{"#":"0.a_i32","D":3}]},"N":"perm"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_9/MultiHeadAttention/"},"N":"struct_name"}],"I":[{"%":682}],"O":[{"%":683,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,12,512,64],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.full","A":[{"AT":{"#":"1.a_intarray","D":[1]},"N":"shape"},{"AT":{"#":"0.a_f64","D":0.125},"N":"value"},{"AT":{"#":"1.a_dtype","D":"float32"},"N":"dtype"},{"AT":{"#":"1.a_place","D":[1,0,""]},"N":"place"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_9/MultiHeadAttention/"},"N":"struct_name"}],"I":[],"O":[{"%":684,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[1],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":true}]},"N":"stop_gradient"}]},{"#":"1.scale","A":[{"AT":{"#":"0.a_f32","D":0.0},"N":"bias"},{"AT":{"#":"0.a_bool","D":true},"N":"bias_after_scale"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_9/MultiHeadAttention/"},"N":"struct_name"}],"I":[{"%":673},{"%":684}],"O":[{"%":685,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,12,512,64],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.matmul","A":[{"AT":{"#":"0.a_bool","D":false},"N":"transpose_x"},{"AT":{"#":"0.a_bool","D":true},"N":"transpose_y"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_9/MultiHeadAttention/"},"N":"struct_name"}],"I":[{"%":685},{"%":680}],"O":[{"%":686,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,12,512,512],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.add","A":[{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_9/MultiHeadAttention/"},"N":"struct_name"}],"I":[{"%":686},{"%":212}],"O":[{"%":687,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,12,512,512],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.softmax","A":[{"AT":{"#":"0.a_i32","D":-1},"N":"axis"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_9/MultiHeadAttention/"},"N":"struct_name"}],"I":[{"%":687}],"O":[{"%":688,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,12,512,512],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.full","A":[{"AT":{"#":"1.a_intarray","D":[1]},"N":"shape"},{"AT":{"#":"0.a_f64","D":0.10000000149011612},"N":"value"},{"AT":{"#":"1.a_dtype","D":"float32"},"N":"dtype"},{"AT":{"#":"1.a_place","D":[1,0,""]},"N":"place"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_9/MultiHeadAttention/"},"N":"struct_name"}],"I":[],"O":[{"%":689,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[1],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":true}]},"N":"stop_gradient"}]},{"#":"1.dropout","A":[{"AT":{"#":"0.a_bool","D":true},"N":"is_test"},{"AT":{"#":"0.a_str","D":"upscale_in_train"},"N":"mode"},{"AT":{"#":"0.a_i32","D":0},"N":"seed"},{"AT":{"#":"0.a_bool","D":false},"N":"fix_seed"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_9/MultiHeadAttention/"},"N":"struct_name"}],"I":[{"%":688},{"%":0},{"%":689}],"O":[{"%":690,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,12,512,512],"NCHW",[],0]}},{"%":691,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_ui8"},[-1,12,512,512],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false},{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.matmul","A":[{"AT":{"#":"0.a_bool","D":false},"N":"transpose_x"},{"AT":{"#":"0.a_bool","D":false},"N":"transpose_y"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_9/MultiHeadAttention/"},"N":"struct_name"}],"I":[{"%":690},{"%":683}],"O":[{"%":692,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,12,512,64],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.transpose","A":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_i32","D":0},{"#":"0.a_i32","D":2},{"#":"0.a_i32","D":1},{"#":"0.a_i32","D":3}]},"N":"perm"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_9/MultiHeadAttention/"},"N":"struct_name"}],"I":[{"%":692}],"O":[{"%":693,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512,12,64],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.full_int_array","A":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_i64","D":0},{"#":"0.a_i64","D":0},{"#":"0.a_i64","D":768}]},"N":"value"},{"AT":{"#":"1.a_dtype","D":"int64"},"N":"dtype"},{"AT":{"#":"1.a_place","D":[1,0,""]},"N":"place"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_9/MultiHeadAttention/"},"N":"struct_name"}],"I":[],"O":[{"%":694,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_i64"},[3],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":true}]},"N":"stop_gradient"}]},{"#":"1.reshape","A":[{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_9/MultiHeadAttention/"},"N":"struct_name"}],"I":[{"%":693},{"%":694}],"O":[{"%":695,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512,768],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.matmul","A":[{"AT":{"#":"0.a_bool","D":false},"N":"transpose_x"},{"AT":{"#":"0.a_bool","D":false},"N":"transpose_y"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_9/MultiHeadAttention/Linear_3/"},"N":"struct_name"}],"I":[{"%":695},{"%":46}],"O":[{"%":696,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512,768],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.add","A":[{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_9/MultiHeadAttention/Linear_3/"},"N":"struct_name"}],"I":[{"%":696},{"%":45}],"O":[{"%":697,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512,768],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.full","A":[{"AT":{"#":"1.a_intarray","D":[1]},"N":"shape"},{"AT":{"#":"0.a_f64","D":0.10000000149011612},"N":"value"},{"AT":{"#":"1.a_dtype","D":"float32"},"N":"dtype"},{"AT":{"#":"1.a_place","D":[1,0,""]},"N":"place"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_9/Dropout/"},"N":"struct_name"}],"I":[],"O":[{"%":698,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[1],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":true}]},"N":"stop_gradient"}]},{"#":"1.dropout","A":[{"AT":{"#":"0.a_bool","D":true},"N":"is_test"},{"AT":{"#":"0.a_str","D":"upscale_in_train"},"N":"mode"},{"AT":{"#":"0.a_i32","D":0},"N":"seed"},{"AT":{"#":"0.a_bool","D":false},"N":"fix_seed"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_9/Dropout/"},"N":"struct_name"}],"I":[{"%":697},{"%":0},{"%":698}],"O":[{"%":699,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512,768],"NCHW",[],0]}},{"%":700,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_ui8"},[-1,512,768],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false},{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.add","A":[{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_9/"},"N":"struct_name"}],"I":[{"%":666},{"%":699}],"O":[{"%":701,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512,768],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.layer_norm","A":[{"AT":{"#":"0.a_f32","D":9.999999960041972e-13},"N":"epsilon"},{"AT":{"#":"0.a_i32","D":2},"N":"begin_norm_axis"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_9/LayerNorm/"},"N":"struct_name"}],"I":[{"%":701},{"%":40},{"%":39}],"O":[{"%":702,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512,768],"NCHW",[],0]}},{"%":703,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512],"NCHW",[],0]}},{"%":704,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false},{"#":"0.a_bool","D":false},{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.matmul","A":[{"AT":{"#":"0.a_bool","D":false},"N":"transpose_x"},{"AT":{"#":"0.a_bool","D":false},"N":"transpose_y"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_9/Linear/"},"N":"struct_name"}],"I":[{"%":702},{"%":44}],"O":[{"%":705,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512,3072],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.add","A":[{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_9/Linear/"},"N":"struct_name"}],"I":[{"%":705},{"%":43}],"O":[{"%":706,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512,3072],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.gelu","A":[{"AT":{"#":"0.a_bool","D":false},"N":"approximate"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_9/"},"N":"struct_name"}],"I":[{"%":706}],"O":[{"%":707,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512,3072],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.matmul","A":[{"AT":{"#":"0.a_bool","D":false},"N":"transpose_x"},{"AT":{"#":"0.a_bool","D":false},"N":"transpose_y"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_9/Linear_1/"},"N":"struct_name"}],"I":[{"%":707},{"%":42}],"O":[{"%":708,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512,768],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.add","A":[{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_9/Linear_1/"},"N":"struct_name"}],"I":[{"%":708},{"%":41}],"O":[{"%":709,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512,768],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.full","A":[{"AT":{"#":"1.a_intarray","D":[1]},"N":"shape"},{"AT":{"#":"0.a_f64","D":0.10000000149011612},"N":"value"},{"AT":{"#":"1.a_dtype","D":"float32"},"N":"dtype"},{"AT":{"#":"1.a_place","D":[1,0,""]},"N":"place"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_9/Dropout_2/"},"N":"struct_name"}],"I":[],"O":[{"%":710,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[1],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":true}]},"N":"stop_gradient"}]},{"#":"1.dropout","A":[{"AT":{"#":"0.a_bool","D":true},"N":"is_test"},{"AT":{"#":"0.a_str","D":"upscale_in_train"},"N":"mode"},{"AT":{"#":"0.a_i32","D":0},"N":"seed"},{"AT":{"#":"0.a_bool","D":false},"N":"fix_seed"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_9/Dropout_2/"},"N":"struct_name"}],"I":[{"%":709},{"%":0},{"%":710}],"O":[{"%":711,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512,768],"NCHW",[],0]}},{"%":712,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_ui8"},[-1,512,768],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false},{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.add","A":[{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_9/"},"N":"struct_name"}],"I":[{"%":702},{"%":711}],"O":[{"%":713,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512,768],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.layer_norm","A":[{"AT":{"#":"0.a_f32","D":9.999999960041972e-13},"N":"epsilon"},{"AT":{"#":"0.a_i32","D":2},"N":"begin_norm_axis"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_9/LayerNorm_1/"},"N":"struct_name"}],"I":[{"%":713},{"%":38},{"%":37}],"O":[{"%":714,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512,768],"NCHW",[],0]}},{"%":715,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512],"NCHW",[],0]}},{"%":716,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false},{"#":"0.a_bool","D":false},{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.matmul","A":[{"AT":{"#":"0.a_bool","D":false},"N":"transpose_x"},{"AT":{"#":"0.a_bool","D":false},"N":"transpose_y"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_10/MultiHeadAttention/Linear/"},"N":"struct_name"}],"I":[{"%":714},{"%":36}],"O":[{"%":717,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512,768],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.add","A":[{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_10/MultiHeadAttention/Linear/"},"N":"struct_name"}],"I":[{"%":717},{"%":35}],"O":[{"%":718,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512,768],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.full_int_array","A":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_i64","D":0},{"#":"0.a_i64","D":0},{"#":"0.a_i64","D":12},{"#":"0.a_i64","D":64}]},"N":"value"},{"AT":{"#":"1.a_dtype","D":"int64"},"N":"dtype"},{"AT":{"#":"1.a_place","D":[1,0,""]},"N":"place"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_10/MultiHeadAttention/"},"N":"struct_name"}],"I":[],"O":[{"%":719,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_i64"},[4],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":true}]},"N":"stop_gradient"}]},{"#":"1.reshape","A":[{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_10/MultiHeadAttention/"},"N":"struct_name"}],"I":[{"%":718},{"%":719}],"O":[{"%":720,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512,12,64],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.transpose","A":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_i32","D":0},{"#":"0.a_i32","D":2},{"#":"0.a_i32","D":1},{"#":"0.a_i32","D":3}]},"N":"perm"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_10/MultiHeadAttention/"},"N":"struct_name"}],"I":[{"%":720}],"O":[{"%":721,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,12,512,64],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.matmul","A":[{"AT":{"#":"0.a_bool","D":false},"N":"transpose_x"},{"AT":{"#":"0.a_bool","D":false},"N":"transpose_y"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_10/MultiHeadAttention/Linear_1/"},"N":"struct_name"}],"I":[{"%":714},{"%":34}],"O":[{"%":722,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512,768],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.add","A":[{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_10/MultiHeadAttention/Linear_1/"},"N":"struct_name"}],"I":[{"%":722},{"%":33}],"O":[{"%":723,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512,768],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.matmul","A":[{"AT":{"#":"0.a_bool","D":false},"N":"transpose_x"},{"AT":{"#":"0.a_bool","D":false},"N":"transpose_y"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_10/MultiHeadAttention/Linear_2/"},"N":"struct_name"}],"I":[{"%":714},{"%":32}],"O":[{"%":724,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512,768],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.add","A":[{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_10/MultiHeadAttention/Linear_2/"},"N":"struct_name"}],"I":[{"%":724},{"%":31}],"O":[{"%":725,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512,768],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.full_int_array","A":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_i64","D":0},{"#":"0.a_i64","D":0},{"#":"0.a_i64","D":12},{"#":"0.a_i64","D":64}]},"N":"value"},{"AT":{"#":"1.a_dtype","D":"int64"},"N":"dtype"},{"AT":{"#":"1.a_place","D":[1,0,""]},"N":"place"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_10/MultiHeadAttention/"},"N":"struct_name"}],"I":[],"O":[{"%":726,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_i64"},[4],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":true}]},"N":"stop_gradient"}]},{"#":"1.reshape","A":[{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_10/MultiHeadAttention/"},"N":"struct_name"}],"I":[{"%":723},{"%":726}],"O":[{"%":727,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512,12,64],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.transpose","A":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_i32","D":0},{"#":"0.a_i32","D":2},{"#":"0.a_i32","D":1},{"#":"0.a_i32","D":3}]},"N":"perm"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_10/MultiHeadAttention/"},"N":"struct_name"}],"I":[{"%":727}],"O":[{"%":728,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,12,512,64],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.full_int_array","A":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_i64","D":0},{"#":"0.a_i64","D":0},{"#":"0.a_i64","D":12},{"#":"0.a_i64","D":64}]},"N":"value"},{"AT":{"#":"1.a_dtype","D":"int64"},"N":"dtype"},{"AT":{"#":"1.a_place","D":[1,0,""]},"N":"place"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_10/MultiHeadAttention/"},"N":"struct_name"}],"I":[],"O":[{"%":729,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_i64"},[4],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":true}]},"N":"stop_gradient"}]},{"#":"1.reshape","A":[{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_10/MultiHeadAttention/"},"N":"struct_name"}],"I":[{"%":725},{"%":729}],"O":[{"%":730,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512,12,64],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.transpose","A":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_i32","D":0},{"#":"0.a_i32","D":2},{"#":"0.a_i32","D":1},{"#":"0.a_i32","D":3}]},"N":"perm"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_10/MultiHeadAttention/"},"N":"struct_name"}],"I":[{"%":730}],"O":[{"%":731,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,12,512,64],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.full","A":[{"AT":{"#":"1.a_intarray","D":[1]},"N":"shape"},{"AT":{"#":"0.a_f64","D":0.125},"N":"value"},{"AT":{"#":"1.a_dtype","D":"float32"},"N":"dtype"},{"AT":{"#":"1.a_place","D":[1,0,""]},"N":"place"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_10/MultiHeadAttention/"},"N":"struct_name"}],"I":[],"O":[{"%":732,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[1],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":true}]},"N":"stop_gradient"}]},{"#":"1.scale","A":[{"AT":{"#":"0.a_f32","D":0.0},"N":"bias"},{"AT":{"#":"0.a_bool","D":true},"N":"bias_after_scale"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_10/MultiHeadAttention/"},"N":"struct_name"}],"I":[{"%":721},{"%":732}],"O":[{"%":733,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,12,512,64],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.matmul","A":[{"AT":{"#":"0.a_bool","D":false},"N":"transpose_x"},{"AT":{"#":"0.a_bool","D":true},"N":"transpose_y"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_10/MultiHeadAttention/"},"N":"struct_name"}],"I":[{"%":733},{"%":728}],"O":[{"%":734,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,12,512,512],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.add","A":[{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_10/MultiHeadAttention/"},"N":"struct_name"}],"I":[{"%":734},{"%":212}],"O":[{"%":735,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,12,512,512],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.softmax","A":[{"AT":{"#":"0.a_i32","D":-1},"N":"axis"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_10/MultiHeadAttention/"},"N":"struct_name"}],"I":[{"%":735}],"O":[{"%":736,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,12,512,512],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.full","A":[{"AT":{"#":"1.a_intarray","D":[1]},"N":"shape"},{"AT":{"#":"0.a_f64","D":0.10000000149011612},"N":"value"},{"AT":{"#":"1.a_dtype","D":"float32"},"N":"dtype"},{"AT":{"#":"1.a_place","D":[1,0,""]},"N":"place"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_10/MultiHeadAttention/"},"N":"struct_name"}],"I":[],"O":[{"%":737,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[1],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":true}]},"N":"stop_gradient"}]},{"#":"1.dropout","A":[{"AT":{"#":"0.a_bool","D":true},"N":"is_test"},{"AT":{"#":"0.a_str","D":"upscale_in_train"},"N":"mode"},{"AT":{"#":"0.a_i32","D":0},"N":"seed"},{"AT":{"#":"0.a_bool","D":false},"N":"fix_seed"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_10/MultiHeadAttention/"},"N":"struct_name"}],"I":[{"%":736},{"%":0},{"%":737}],"O":[{"%":738,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,12,512,512],"NCHW",[],0]}},{"%":739,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_ui8"},[-1,12,512,512],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false},{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.matmul","A":[{"AT":{"#":"0.a_bool","D":false},"N":"transpose_x"},{"AT":{"#":"0.a_bool","D":false},"N":"transpose_y"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_10/MultiHeadAttention/"},"N":"struct_name"}],"I":[{"%":738},{"%":731}],"O":[{"%":740,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,12,512,64],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.transpose","A":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_i32","D":0},{"#":"0.a_i32","D":2},{"#":"0.a_i32","D":1},{"#":"0.a_i32","D":3}]},"N":"perm"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_10/MultiHeadAttention/"},"N":"struct_name"}],"I":[{"%":740}],"O":[{"%":741,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512,12,64],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.full_int_array","A":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_i64","D":0},{"#":"0.a_i64","D":0},{"#":"0.a_i64","D":768}]},"N":"value"},{"AT":{"#":"1.a_dtype","D":"int64"},"N":"dtype"},{"AT":{"#":"1.a_place","D":[1,0,""]},"N":"place"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_10/MultiHeadAttention/"},"N":"struct_name"}],"I":[],"O":[{"%":742,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_i64"},[3],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":true}]},"N":"stop_gradient"}]},{"#":"1.reshape","A":[{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_10/MultiHeadAttention/"},"N":"struct_name"}],"I":[{"%":741},{"%":742}],"O":[{"%":743,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512,768],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.matmul","A":[{"AT":{"#":"0.a_bool","D":false},"N":"transpose_x"},{"AT":{"#":"0.a_bool","D":false},"N":"transpose_y"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_10/MultiHeadAttention/Linear_3/"},"N":"struct_name"}],"I":[{"%":743},{"%":30}],"O":[{"%":744,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512,768],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.add","A":[{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_10/MultiHeadAttention/Linear_3/"},"N":"struct_name"}],"I":[{"%":744},{"%":29}],"O":[{"%":745,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512,768],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.full","A":[{"AT":{"#":"1.a_intarray","D":[1]},"N":"shape"},{"AT":{"#":"0.a_f64","D":0.10000000149011612},"N":"value"},{"AT":{"#":"1.a_dtype","D":"float32"},"N":"dtype"},{"AT":{"#":"1.a_place","D":[1,0,""]},"N":"place"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_10/Dropout/"},"N":"struct_name"}],"I":[],"O":[{"%":746,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[1],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":true}]},"N":"stop_gradient"}]},{"#":"1.dropout","A":[{"AT":{"#":"0.a_bool","D":true},"N":"is_test"},{"AT":{"#":"0.a_str","D":"upscale_in_train"},"N":"mode"},{"AT":{"#":"0.a_i32","D":0},"N":"seed"},{"AT":{"#":"0.a_bool","D":false},"N":"fix_seed"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_10/Dropout/"},"N":"struct_name"}],"I":[{"%":745},{"%":0},{"%":746}],"O":[{"%":747,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512,768],"NCHW",[],0]}},{"%":748,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_ui8"},[-1,512,768],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false},{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.add","A":[{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_10/"},"N":"struct_name"}],"I":[{"%":714},{"%":747}],"O":[{"%":749,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512,768],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.layer_norm","A":[{"AT":{"#":"0.a_f32","D":9.999999960041972e-13},"N":"epsilon"},{"AT":{"#":"0.a_i32","D":2},"N":"begin_norm_axis"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_10/LayerNorm/"},"N":"struct_name"}],"I":[{"%":749},{"%":24},{"%":23}],"O":[{"%":750,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512,768],"NCHW",[],0]}},{"%":751,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512],"NCHW",[],0]}},{"%":752,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false},{"#":"0.a_bool","D":false},{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.matmul","A":[{"AT":{"#":"0.a_bool","D":false},"N":"transpose_x"},{"AT":{"#":"0.a_bool","D":false},"N":"transpose_y"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_10/Linear/"},"N":"struct_name"}],"I":[{"%":750},{"%":28}],"O":[{"%":753,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512,3072],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.add","A":[{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_10/Linear/"},"N":"struct_name"}],"I":[{"%":753},{"%":27}],"O":[{"%":754,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512,3072],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.gelu","A":[{"AT":{"#":"0.a_bool","D":false},"N":"approximate"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_10/"},"N":"struct_name"}],"I":[{"%":754}],"O":[{"%":755,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512,3072],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.matmul","A":[{"AT":{"#":"0.a_bool","D":false},"N":"transpose_x"},{"AT":{"#":"0.a_bool","D":false},"N":"transpose_y"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_10/Linear_1/"},"N":"struct_name"}],"I":[{"%":755},{"%":26}],"O":[{"%":756,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512,768],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.add","A":[{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_10/Linear_1/"},"N":"struct_name"}],"I":[{"%":756},{"%":25}],"O":[{"%":757,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512,768],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.full","A":[{"AT":{"#":"1.a_intarray","D":[1]},"N":"shape"},{"AT":{"#":"0.a_f64","D":0.10000000149011612},"N":"value"},{"AT":{"#":"1.a_dtype","D":"float32"},"N":"dtype"},{"AT":{"#":"1.a_place","D":[1,0,""]},"N":"place"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_10/Dropout_2/"},"N":"struct_name"}],"I":[],"O":[{"%":758,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[1],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":true}]},"N":"stop_gradient"}]},{"#":"1.dropout","A":[{"AT":{"#":"0.a_bool","D":true},"N":"is_test"},{"AT":{"#":"0.a_str","D":"upscale_in_train"},"N":"mode"},{"AT":{"#":"0.a_i32","D":0},"N":"seed"},{"AT":{"#":"0.a_bool","D":false},"N":"fix_seed"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_10/Dropout_2/"},"N":"struct_name"}],"I":[{"%":757},{"%":0},{"%":758}],"O":[{"%":759,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512,768],"NCHW",[],0]}},{"%":760,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_ui8"},[-1,512,768],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false},{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.add","A":[{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_10/"},"N":"struct_name"}],"I":[{"%":750},{"%":759}],"O":[{"%":761,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512,768],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.layer_norm","A":[{"AT":{"#":"0.a_f32","D":9.999999960041972e-13},"N":"epsilon"},{"AT":{"#":"0.a_i32","D":2},"N":"begin_norm_axis"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_10/LayerNorm_1/"},"N":"struct_name"}],"I":[{"%":761},{"%":22},{"%":21}],"O":[{"%":762,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512,768],"NCHW",[],0]}},{"%":763,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512],"NCHW",[],0]}},{"%":764,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false},{"#":"0.a_bool","D":false},{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.matmul","A":[{"AT":{"#":"0.a_bool","D":false},"N":"transpose_x"},{"AT":{"#":"0.a_bool","D":false},"N":"transpose_y"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_11/MultiHeadAttention/Linear/"},"N":"struct_name"}],"I":[{"%":762},{"%":20}],"O":[{"%":765,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512,768],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.add","A":[{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_11/MultiHeadAttention/Linear/"},"N":"struct_name"}],"I":[{"%":765},{"%":19}],"O":[{"%":766,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512,768],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.full_int_array","A":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_i64","D":0},{"#":"0.a_i64","D":0},{"#":"0.a_i64","D":12},{"#":"0.a_i64","D":64}]},"N":"value"},{"AT":{"#":"1.a_dtype","D":"int64"},"N":"dtype"},{"AT":{"#":"1.a_place","D":[1,0,""]},"N":"place"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_11/MultiHeadAttention/"},"N":"struct_name"}],"I":[],"O":[{"%":767,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_i64"},[4],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":true}]},"N":"stop_gradient"}]},{"#":"1.reshape","A":[{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_11/MultiHeadAttention/"},"N":"struct_name"}],"I":[{"%":766},{"%":767}],"O":[{"%":768,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512,12,64],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.transpose","A":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_i32","D":0},{"#":"0.a_i32","D":2},{"#":"0.a_i32","D":1},{"#":"0.a_i32","D":3}]},"N":"perm"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_11/MultiHeadAttention/"},"N":"struct_name"}],"I":[{"%":768}],"O":[{"%":769,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,12,512,64],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.matmul","A":[{"AT":{"#":"0.a_bool","D":false},"N":"transpose_x"},{"AT":{"#":"0.a_bool","D":false},"N":"transpose_y"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_11/MultiHeadAttention/Linear_1/"},"N":"struct_name"}],"I":[{"%":762},{"%":18}],"O":[{"%":770,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512,768],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.add","A":[{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_11/MultiHeadAttention/Linear_1/"},"N":"struct_name"}],"I":[{"%":770},{"%":17}],"O":[{"%":771,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512,768],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.matmul","A":[{"AT":{"#":"0.a_bool","D":false},"N":"transpose_x"},{"AT":{"#":"0.a_bool","D":false},"N":"transpose_y"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_11/MultiHeadAttention/Linear_2/"},"N":"struct_name"}],"I":[{"%":762},{"%":16}],"O":[{"%":772,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512,768],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.add","A":[{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_11/MultiHeadAttention/Linear_2/"},"N":"struct_name"}],"I":[{"%":772},{"%":15}],"O":[{"%":773,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512,768],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.full_int_array","A":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_i64","D":0},{"#":"0.a_i64","D":0},{"#":"0.a_i64","D":12},{"#":"0.a_i64","D":64}]},"N":"value"},{"AT":{"#":"1.a_dtype","D":"int64"},"N":"dtype"},{"AT":{"#":"1.a_place","D":[1,0,""]},"N":"place"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_11/MultiHeadAttention/"},"N":"struct_name"}],"I":[],"O":[{"%":774,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_i64"},[4],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":true}]},"N":"stop_gradient"}]},{"#":"1.reshape","A":[{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_11/MultiHeadAttention/"},"N":"struct_name"}],"I":[{"%":771},{"%":774}],"O":[{"%":775,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512,12,64],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.transpose","A":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_i32","D":0},{"#":"0.a_i32","D":2},{"#":"0.a_i32","D":1},{"#":"0.a_i32","D":3}]},"N":"perm"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_11/MultiHeadAttention/"},"N":"struct_name"}],"I":[{"%":775}],"O":[{"%":776,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,12,512,64],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.full_int_array","A":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_i64","D":0},{"#":"0.a_i64","D":0},{"#":"0.a_i64","D":12},{"#":"0.a_i64","D":64}]},"N":"value"},{"AT":{"#":"1.a_dtype","D":"int64"},"N":"dtype"},{"AT":{"#":"1.a_place","D":[1,0,""]},"N":"place"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_11/MultiHeadAttention/"},"N":"struct_name"}],"I":[],"O":[{"%":777,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_i64"},[4],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":true}]},"N":"stop_gradient"}]},{"#":"1.reshape","A":[{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_11/MultiHeadAttention/"},"N":"struct_name"}],"I":[{"%":773},{"%":777}],"O":[{"%":778,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512,12,64],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.transpose","A":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_i32","D":0},{"#":"0.a_i32","D":2},{"#":"0.a_i32","D":1},{"#":"0.a_i32","D":3}]},"N":"perm"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_11/MultiHeadAttention/"},"N":"struct_name"}],"I":[{"%":778}],"O":[{"%":779,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,12,512,64],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.full","A":[{"AT":{"#":"1.a_intarray","D":[1]},"N":"shape"},{"AT":{"#":"0.a_f64","D":0.125},"N":"value"},{"AT":{"#":"1.a_dtype","D":"float32"},"N":"dtype"},{"AT":{"#":"1.a_place","D":[1,0,""]},"N":"place"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_11/MultiHeadAttention/"},"N":"struct_name"}],"I":[],"O":[{"%":780,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[1],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":true}]},"N":"stop_gradient"}]},{"#":"1.scale","A":[{"AT":{"#":"0.a_f32","D":0.0},"N":"bias"},{"AT":{"#":"0.a_bool","D":true},"N":"bias_after_scale"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_11/MultiHeadAttention/"},"N":"struct_name"}],"I":[{"%":769},{"%":780}],"O":[{"%":781,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,12,512,64],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.matmul","A":[{"AT":{"#":"0.a_bool","D":false},"N":"transpose_x"},{"AT":{"#":"0.a_bool","D":true},"N":"transpose_y"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_11/MultiHeadAttention/"},"N":"struct_name"}],"I":[{"%":781},{"%":776}],"O":[{"%":782,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,12,512,512],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.add","A":[{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_11/MultiHeadAttention/"},"N":"struct_name"}],"I":[{"%":782},{"%":212}],"O":[{"%":783,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,12,512,512],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.softmax","A":[{"AT":{"#":"0.a_i32","D":-1},"N":"axis"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_11/MultiHeadAttention/"},"N":"struct_name"}],"I":[{"%":783}],"O":[{"%":784,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,12,512,512],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.full","A":[{"AT":{"#":"1.a_intarray","D":[1]},"N":"shape"},{"AT":{"#":"0.a_f64","D":0.10000000149011612},"N":"value"},{"AT":{"#":"1.a_dtype","D":"float32"},"N":"dtype"},{"AT":{"#":"1.a_place","D":[1,0,""]},"N":"place"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_11/MultiHeadAttention/"},"N":"struct_name"}],"I":[],"O":[{"%":785,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[1],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":true}]},"N":"stop_gradient"}]},{"#":"1.dropout","A":[{"AT":{"#":"0.a_bool","D":true},"N":"is_test"},{"AT":{"#":"0.a_str","D":"upscale_in_train"},"N":"mode"},{"AT":{"#":"0.a_i32","D":0},"N":"seed"},{"AT":{"#":"0.a_bool","D":false},"N":"fix_seed"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_11/MultiHeadAttention/"},"N":"struct_name"}],"I":[{"%":784},{"%":0},{"%":785}],"O":[{"%":786,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,12,512,512],"NCHW",[],0]}},{"%":787,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_ui8"},[-1,12,512,512],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false},{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.matmul","A":[{"AT":{"#":"0.a_bool","D":false},"N":"transpose_x"},{"AT":{"#":"0.a_bool","D":false},"N":"transpose_y"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_11/MultiHeadAttention/"},"N":"struct_name"}],"I":[{"%":786},{"%":779}],"O":[{"%":788,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,12,512,64],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.transpose","A":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_i32","D":0},{"#":"0.a_i32","D":2},{"#":"0.a_i32","D":1},{"#":"0.a_i32","D":3}]},"N":"perm"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_11/MultiHeadAttention/"},"N":"struct_name"}],"I":[{"%":788}],"O":[{"%":789,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512,12,64],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.full_int_array","A":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_i64","D":0},{"#":"0.a_i64","D":0},{"#":"0.a_i64","D":768}]},"N":"value"},{"AT":{"#":"1.a_dtype","D":"int64"},"N":"dtype"},{"AT":{"#":"1.a_place","D":[1,0,""]},"N":"place"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_11/MultiHeadAttention/"},"N":"struct_name"}],"I":[],"O":[{"%":790,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_i64"},[3],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":true}]},"N":"stop_gradient"}]},{"#":"1.reshape","A":[{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_11/MultiHeadAttention/"},"N":"struct_name"}],"I":[{"%":789},{"%":790}],"O":[{"%":791,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512,768],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.matmul","A":[{"AT":{"#":"0.a_bool","D":false},"N":"transpose_x"},{"AT":{"#":"0.a_bool","D":false},"N":"transpose_y"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_11/MultiHeadAttention/Linear_3/"},"N":"struct_name"}],"I":[{"%":791},{"%":14}],"O":[{"%":792,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512,768],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.add","A":[{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_11/MultiHeadAttention/Linear_3/"},"N":"struct_name"}],"I":[{"%":792},{"%":13}],"O":[{"%":793,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512,768],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.full","A":[{"AT":{"#":"1.a_intarray","D":[1]},"N":"shape"},{"AT":{"#":"0.a_f64","D":0.10000000149011612},"N":"value"},{"AT":{"#":"1.a_dtype","D":"float32"},"N":"dtype"},{"AT":{"#":"1.a_place","D":[1,0,""]},"N":"place"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_11/Dropout/"},"N":"struct_name"}],"I":[],"O":[{"%":794,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[1],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":true}]},"N":"stop_gradient"}]},{"#":"1.dropout","A":[{"AT":{"#":"0.a_bool","D":true},"N":"is_test"},{"AT":{"#":"0.a_str","D":"upscale_in_train"},"N":"mode"},{"AT":{"#":"0.a_i32","D":0},"N":"seed"},{"AT":{"#":"0.a_bool","D":false},"N":"fix_seed"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_11/Dropout/"},"N":"struct_name"}],"I":[{"%":793},{"%":0},{"%":794}],"O":[{"%":795,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512,768],"NCHW",[],0]}},{"%":796,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_ui8"},[-1,512,768],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false},{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.add","A":[{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_11/"},"N":"struct_name"}],"I":[{"%":762},{"%":795}],"O":[{"%":797,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512,768],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.layer_norm","A":[{"AT":{"#":"0.a_f32","D":9.999999960041972e-13},"N":"epsilon"},{"AT":{"#":"0.a_i32","D":2},"N":"begin_norm_axis"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_11/LayerNorm/"},"N":"struct_name"}],"I":[{"%":797},{"%":8},{"%":7}],"O":[{"%":798,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512,768],"NCHW",[],0]}},{"%":799,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512],"NCHW",[],0]}},{"%":800,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false},{"#":"0.a_bool","D":false},{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.matmul","A":[{"AT":{"#":"0.a_bool","D":false},"N":"transpose_x"},{"AT":{"#":"0.a_bool","D":false},"N":"transpose_y"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_11/Linear/"},"N":"struct_name"}],"I":[{"%":798},{"%":12}],"O":[{"%":801,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512,3072],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.add","A":[{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_11/Linear/"},"N":"struct_name"}],"I":[{"%":801},{"%":11}],"O":[{"%":802,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512,3072],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.gelu","A":[{"AT":{"#":"0.a_bool","D":false},"N":"approximate"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_11/"},"N":"struct_name"}],"I":[{"%":802}],"O":[{"%":803,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512,3072],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.matmul","A":[{"AT":{"#":"0.a_bool","D":false},"N":"transpose_x"},{"AT":{"#":"0.a_bool","D":false},"N":"transpose_y"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_11/Linear_1/"},"N":"struct_name"}],"I":[{"%":803},{"%":10}],"O":[{"%":804,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512,768],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.add","A":[{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_11/Linear_1/"},"N":"struct_name"}],"I":[{"%":804},{"%":9}],"O":[{"%":805,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512,768],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.full","A":[{"AT":{"#":"1.a_intarray","D":[1]},"N":"shape"},{"AT":{"#":"0.a_f64","D":0.10000000149011612},"N":"value"},{"AT":{"#":"1.a_dtype","D":"float32"},"N":"dtype"},{"AT":{"#":"1.a_place","D":[1,0,""]},"N":"place"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_11/Dropout_2/"},"N":"struct_name"}],"I":[],"O":[{"%":806,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[1],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":true}]},"N":"stop_gradient"}]},{"#":"1.dropout","A":[{"AT":{"#":"0.a_bool","D":true},"N":"is_test"},{"AT":{"#":"0.a_str","D":"upscale_in_train"},"N":"mode"},{"AT":{"#":"0.a_i32","D":0},"N":"seed"},{"AT":{"#":"0.a_bool","D":false},"N":"fix_seed"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_11/Dropout_2/"},"N":"struct_name"}],"I":[{"%":805},{"%":0},{"%":806}],"O":[{"%":807,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512,768],"NCHW",[],0]}},{"%":808,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_ui8"},[-1,512,768],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false},{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.add","A":[{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_11/"},"N":"struct_name"}],"I":[{"%":798},{"%":807}],"O":[{"%":809,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512,768],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.layer_norm","A":[{"AT":{"#":"0.a_f32","D":9.999999960041972e-13},"N":"epsilon"},{"AT":{"#":"0.a_i32","D":2},"N":"begin_norm_axis"},{"AT":{"#":"0.a_str","D":"/ErnieModel/TransformerEncoder/TransformerEncoderLayer_11/LayerNorm_1/"},"N":"struct_name"}],"I":[{"%":809},{"%":6},{"%":5}],"O":[{"%":810,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512,768],"NCHW",[],0]}},{"%":811,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512],"NCHW",[],0]}},{"%":812,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false},{"#":"0.a_bool","D":false},{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.full","A":[{"AT":{"#":"1.a_intarray","D":[1]},"N":"shape"},{"AT":{"#":"0.a_f64","D":0.10000000149011612},"N":"value"},{"AT":{"#":"1.a_dtype","D":"float32"},"N":"dtype"},{"AT":{"#":"1.a_place","D":[1,0,""]},"N":"place"},{"AT":{"#":"0.a_str","D":"/Dropout/"},"N":"struct_name"}],"I":[],"O":[{"%":813,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[1],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":true}]},"N":"stop_gradient"}]},{"#":"1.dropout","A":[{"AT":{"#":"0.a_bool","D":true},"N":"is_test"},{"AT":{"#":"0.a_str","D":"upscale_in_train"},"N":"mode"},{"AT":{"#":"0.a_i32","D":0},"N":"seed"},{"AT":{"#":"0.a_bool","D":false},"N":"fix_seed"},{"AT":{"#":"0.a_str","D":"/Dropout/"},"N":"struct_name"}],"I":[{"%":810},{"%":0},{"%":813}],"O":[{"%":814,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512,768],"NCHW",[],0]}},{"%":815,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_ui8"},[-1,512,768],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false},{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.matmul","A":[{"AT":{"#":"0.a_bool","D":false},"N":"transpose_x"},{"AT":{"#":"0.a_bool","D":false},"N":"transpose_y"},{"AT":{"#":"0.a_str","D":"/Linear/"},"N":"struct_name"}],"I":[{"%":814},{"%":2}],"O":[{"%":816,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512,21],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.add","A":[{"AT":{"#":"0.a_str","D":"/Linear/"},"N":"struct_name"}],"I":[{"%":816},{"%":1}],"O":[{"%":817,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512,21],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"}]},{"#":"1.fetch","A":[{"AT":{"#":"0.a_str","D":"fetch_name_0"},"N":"name"},{"AT":{"#":"0.a_i32","D":0},"N":"col"}],"I":[{"%":817}],"O":[{"%":818,"TT":{"#":"0.t_dtensor","D":[{"#":"0.t_f32"},[-1,512,21],"NCHW",[],0]}}],"OA":[{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":false}]},"N":"stop_gradient"},{"AT":{"#":"0.a_array","D":[{"#":"0.a_bool","D":true}]},"N":"persistable"}]}]}]}]}}