Langchain-Chatchat/requirements.txt

40 lines
1.1 KiB
Plaintext
Raw Normal View History

pymupdf
paddlepaddle==2.4.2
2023-05-21 22:27:02 +08:00
paddleocr~=2.6.1.3
2023-05-20 01:58:08 +08:00
langchain==0.0.174
2023-05-17 13:44:39 +08:00
transformers==4.29.1
2023-04-14 01:06:13 +08:00
unstructured[local-inference]
layoutparser[layoutmodels,tesseract]
2023-05-21 22:27:02 +08:00
nltk~=3.8.1
2023-04-14 01:06:13 +08:00
sentence-transformers
beautifulsoup4
icetk
cpm_kernels
faiss-cpu
gradio==3.28.3
2023-05-21 22:27:02 +08:00
fastapi~=0.95.0
uvicorn~=0.21.1
pypinyin~=0.48.0
click~=8.1.3
2023-05-14 00:36:42 +08:00
tabulate
feedparser
2023-05-21 22:27:02 +08:00
azure-core
2023-05-31 22:31:54 +08:00
openai
2023-05-26 22:32:12 +08:00
#accelerate~=0.18.0
#peft~=0.3.0
#bitsandbytes; platform_system != "Windows"
2023-05-21 22:27:02 +08:00
2023-07-11 19:36:50 +08:00
# 要调用llama-cpp模型如vicuma-13b量化模型需要安装llama-cpp-python库
# but!!! 实测pip install 不好使需要手动从ttps://github.com/abetlen/llama-cpp-python/releases/下载
# 而且注意不同时期的ggml格式并不容!!!因此需要安装的llama-cpp-python版本也不一致需要手动测试才能确定
# 实测ggml-vicuna-13b-1.1在llama-cpp-python 0.1.63上可正常兼容
# 不过本项目模型加载的方式控制的比较严格与llama-cpp-python的兼容性较差很多参数设定不能使用
# 建议如非必要还是不要使用llama-cpp
2023-05-21 22:27:02 +08:00
torch~=2.0.0
pydantic~=1.10.7
starlette~=0.26.1
numpy~=1.23.5
tqdm~=4.65.0
requests~=2.28.2
tenacity~=8.2.2
charset_normalizer==2.1.0