tocmo0nlord/axolotl
1
0
Fork 0
Code Issues Pull Requests Actions 2 Packages Projects Releases Wiki Activity

Prune cuda117 (#327)

Browse Source 
* drop cuda117/torch 1.13.1 from support, pin flash attention to v2.0.1, rm torchvision/torchaudio install

* gptq base build not needed. add sm 9.0 support
This commit is contained in:
Wing Lian
2023-07-26 16:27:49 -04:00
committed by GitHub
parent 9f69c4d8c1
commit 2c37bf6c21
3 changed files with 12 additions and 29 deletions
Download Patch File Download Diff File Expand all files Collapse all files

16
.github/workflows/base.yml vendored
View File

@@ -19,22 +19,12 @@ jobs:
cuda_version: 11.8.0
python_version: "3.9"
pytorch: 2.0.1
axolotl_extras:
torch_cuda_arch_list: "7.0 7.5 8.0 8.6 9.0+PTX"
- cuda: "118"
cuda_version: 11.8.0
python_version: "3.10"
pytorch: 2.0.1
axolotl_extras:
- cuda: "117"
cuda_version: 11.7.1
python_version: "3.9"
pytorch: 1.13.1
axolotl_extras:
- cuda: "118"
cuda_version: 11.8.0
python_version: "3.9"
pytorch: 2.0.1
axolotl_extras: gptq
torch_cuda_arch_list: "7.0 7.5 8.0 8.6 9.0+PTX"
steps:
- name: Checkout
uses: actions/checkout@v3
@@ -63,4 +53,4 @@ jobs:
CUDA=${{ matrix.cuda }}
PYTHON_VERSION=${{ matrix.python_version }}
PYTORCH_VERSION=${{ matrix.pytorch }}
AXOLOTL_EXTRAS=${{ matrix.axolotl_extras }}
TORCH_CUDA_ARCH_LIST=${{ matrix.torch_cuda_arch_list }}

12
.github/workflows/main.yml vendored
View File

@@ -29,11 +29,6 @@ jobs:
python_version: "3.9"
pytorch: 2.0.1
axolotl_extras: gptq
- cuda: cu117
cuda_version: 11.7.1
python_version: "3.9"
pytorch: 1.13.1
axolotl_extras:
runs-on: self-hosted
steps:
- name: Checkout
@@ -55,7 +50,7 @@ jobs:
with:
context: .
build-args: |
BASE_TAG=${{ github.ref_name }}-base-py${{ matrix.python_version }}-${{ matrix.cuda }}-${{ matrix.pytorch }}${{ matrix.axolotl_extras != '' && '-' || '' }}${{ matrix.axolotl_extras }}
BASE_TAG=${{ github.ref_name }}-base-py${{ matrix.python_version }}-${{ matrix.cuda }}-${{ matrix.pytorch }}
file: ./docker/Dockerfile
push: ${{ github.event_name != 'pull_request' }}
tags: ${{ steps.metadata.outputs.tags }}-py${{ matrix.python_version }}-${{ matrix.cuda }}-${{ matrix.pytorch }}${{ matrix.axolotl_extras != '' && '-' || '' }}${{ matrix.axolotl_extras }}
@@ -82,11 +77,6 @@ jobs:
python_version: "3.9"
pytorch: 2.0.1
axolotl_extras: gptq
- cuda: 117
cuda_version: 11.7.1
python_version: "3.9"
pytorch: 1.13.1
axolotl_extras:
runs-on: self-hosted
steps:
- name: Checkout

13
docker/Dockerfile-base
View File

@@ -8,7 +8,7 @@ FROM nvidia/cuda:$CUDA_VERSION-cudnn$CUDNN_VERSION-devel-ubuntu$UBUNTU_VERSION a
ENV PATH="/root/miniconda3/bin:${PATH}"
ARG PYTHON_VERSION="3.9"
ARG PYTORCH="2.0.0"
ARG PYTORCH_VERSION="2.0.1"
ARG CUDA="118"
ENV PYTHON_VERSION=$PYTHON_VERSION
@@ -29,18 +29,18 @@ ENV PATH="/root/miniconda3/envs/py${PYTHON_VERSION}/bin:${PATH}"
WORKDIR /workspace
RUN python3 -m pip install --upgrade pip && pip3 install packaging && \
python3 -m pip install --no-cache-dir -U torch==${PYTORCH} torchvision torchaudio --extra-index-url https://download.pytorch.org/whl/cu$CUDA
python3 -m pip install --no-cache-dir -U torch==${PYTORCH_VERSION}+cu${CUDA} --extra-index-url https://download.pytorch.org/whl/cu$CUDA
FROM base-builder AS flash-attn-builder
WORKDIR /workspace
ARG TORCH_CUDA_ARCH_LIST="7.0 7.5 8.0 8.6+PTX"
ARG TORCH_CUDA_ARCH_LIST="7.0 7.5 8.0 8.6 9.0+PTX"
RUN git clone https://github.com/Dao-AILab/flash-attention.git && \
cd flash-attention && \
git checkout 9ee0ff1 && \
git checkout v2.0.1 && \
python3 setup.py bdist_wheel && \
cd csrc/fused_dense_lib && \
python3 setup.py bdist_wheel && \
@@ -53,7 +53,7 @@ RUN git clone https://github.com/Dao-AILab/flash-attention.git && \
FROM base-builder AS deepspeed-builder
ARG TORCH_CUDA_ARCH_LIST="7.0 7.5 8.0 8.6+PTX"
ARG TORCH_CUDA_ARCH_LIST="7.0 7.5 8.0 8.6 9.0+PTX"
WORKDIR /workspace
@@ -74,6 +74,9 @@ RUN git clone https://github.com/TimDettmers/bitsandbytes.git && \
FROM base-builder
ARG TORCH_CUDA_ARCH_LIST="7.0 7.5 8.0 8.6 9.0+PTX"
ENV TORCH_CUDA_ARCH_LIST=$TORCH_CUDA_ARCH_LIST
# recompile apex
RUN python3 -m pip uninstall -y apex
RUN git clone https://github.com/NVIDIA/apex