qwen-7b-chat / requirements.txt
ffreemt
Update bitsandbytes device_map={'': 0} rich, check nvidia-smi with subproces.run
b55d7c2
# https://www.zhihu.com/question/309583980
# cudatoolkit # this is for conda only
# transformers==4.31.0
transformers # ==4.30.2
accelerate
tiktoken
einops
# flash-attention
# git clone -b v1.0.8 https://github.com/Dao-AILab/flash-attention
# cd flash-attention && pip install .
# pip install csrc/layer_norm
# pip install csrc/rotary
torch==2.0.1 # 2.0.1
safetensors
# bitsandbytes
# git+https://github.com/TimDettmers/bitsandbytes.git
# bitsandbytes==0.39.0
# bitsandbytes-cuda113 # This version of bitsandbytes is deprecated. Please switch to `pip install bitsandbytes`
bitsandbytes
transformers_stream_generator
scipy
loguru
about-time
rich
# mpmath, lit, cmake, bitsandbytes, sympy, scipy, safetensors, regex, pygments, psutil, nvidia-nvtx-cu11, nvidia-nccl-cu11, nvidia-cusparse-cu11, nvidia-curand-cu11, nvidia-cufft-cu11, nvidia-cuda-runtime-cu11, nvidia-cuda-nvrtc-cu11, nvidia-cuda-cupti-cu11, nvidia-cublas-cu11, networkx, mdurl, MarkupSafe, loguru, einops, about-time, tiktoken, nvidia-cusolver-cu11, nvidia-cudnn-cu11, markdown-it-py, jinja2, tokenizers, rich, transformers, transformers_stream_generator, triton, torch, accelerate
# Successfully installed aiofiles-23.2.1 altair-5.1.2 anyio-3.7.1 contourpy-1.1.1 cycler-0.12.1 exceptiongroup-1.1.3 fastapi-0.103.2 ffmpy-0.3.1 fonttools-4.43.1 gradio-3.39.0 gradio-client-0.6.0 httpcore-0.18.0 httpx-0.25.0 jsonschema-4.19.1 jsonschema-specifications-2023.7.1 kiwisolver-1.4.5 linkify-it-py-2.0.2 markdown-it-py-2.2.0 matplotlib-3.8.0 mdit-py-plugins-0.3.3 orjson-3.9.7 pillow-10.0.1 pydub-0.25.1 pyparsing-3.1.1 python-multipart-0.0.6 referencing-0.30.2 rpds-py-0.10.4 semantic-version-2.10.0 sniffio-1.3.0 spaces-0.16.1 starlette-0.27.0 toolz-0.12.0 uc-micro-py-1.0.2 websockets-11.0.3