huggingface/text-generation-inference
5
0
Fork 0
mirror of https://github.com/huggingface/text-generation-inference.git synced 2025-09-11 04:14:52 +00:00
Code Issues Packages Projects Releases Wiki Activity

...

Browse Source
This commit is contained in:
Nicolas Patry 2025-02-07 17:01:48 +01:00
parent 0cd364f313
commit d743252d44
No known key found for this signature in database
GPG Key ID: 4242CEF24CB6DBF9
1 changed files with 3 additions and 2 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

5
Dockerfile
View File

@ -209,14 +209,15 @@ ENV UV_SYSTEM_PYTHON=1
ENV HF_KERNELS_CACHE=/kernels
RUN cd server && \
pip install -U pip uv && \
uv sync --frozen --extra gen --extra attention --extra bnb --extra accelerate --extra compressed-tensors --extra marlin --extra moe --extra quantize --extra peft --extra outlines --extra nvidia-nccl-cu12 --no-install-project && \
uv sync --frozen --extra gen --extra attention --extra bnb --extra accelerate --extra compressed-tensors --extra marlin --extra moe --extra quantize --extra peft --extra outlines --no-install-project && \
. ./.venv/bin/activate && \
make gen-server-raw && \
hf-kernels download .
RUN cd server && \
uv sync --frozen --extra gen --extra attention --extra bnb --extra accelerate --extra compressed-tensors --extra marlin --extra moe --extra quantize --extra peft --extra outlines --extra nvidia-nccl-cu12 && \
uv sync --frozen --extra gen --extra attention --extra bnb --extra accelerate --extra compressed-tensors --extra marlin --extra moe --extra quantize --extra peft --extra outlines && \
. ./.venv/bin/activate && \
uv pip install nvidia-nccl-cu12==2.22.3 && \
pwd && \
text-generation-server --help