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

add cu130 + torch 2.9.1 to test matrices (#3343)

Browse Source 
* add cu130 + torch 2.9.1 to test matrices

* uv can't use pip3 directly
This commit is contained in:
Wing Lian
2026-01-05 15:24:29 -05:00
committed by GitHub
parent 4e61b8aa23
commit ee59e4de97
3 changed files with 15 additions and 2 deletions
Download Patch File Download Diff File Expand all files Collapse all files

7
.github/workflows/multi-gpu-e2e.yml vendored
View File

@@ -43,6 +43,13 @@ jobs:
axolotl_extras: fbgemm-gpu axolotl_extras: fbgemm-gpu
num_gpus: 2 num_gpus: 2
nightly_build: "true" nightly_build: "true"
- cuda: 130
cuda_version: 13.0.0
python_version: "3.11"
pytorch: 2.9.1
axolotl_extras: fbgemm-gpu
num_gpus: 2
nightly_build: "true"
runs-on: [self-hosted, modal] runs-on: [self-hosted, modal]
timeout-minutes: 120 timeout-minutes: 120
steps: steps:

6
.github/workflows/tests.yml vendored
View File

@@ -316,6 +316,12 @@ jobs:
pytorch: 2.9.1 pytorch: 2.9.1
num_gpus: 1 num_gpus: 1
axolotl_extras: axolotl_extras:
- cuda: 130
cuda_version: 13.0.0
python_version: "3.11"
pytorch: 2.9.1
num_gpus: 1
axolotl_extras:
steps: steps:
- name: Checkout - name: Checkout
uses: actions/checkout@v4 uses: actions/checkout@v4

4
docker/Dockerfile-uv-base
View File

@@ -39,11 +39,11 @@ RUN case "$PYTORCH_VERSION" in \
2.9.[0-9]*) \ 2.9.[0-9]*) \
if [ "$CUDA" = "128" ]; then \ if [ "$CUDA" = "128" ]; then \
wget -nv https://github.com/mjun0812/flash-attention-prebuild-wheels/releases/download/v0.5.4/flash_attn-2.8.3+cu128torch2.9-cp311-cp311-linux_x86_64.whl; \ wget -nv https://github.com/mjun0812/flash-attention-prebuild-wheels/releases/download/v0.5.4/flash_attn-2.8.3+cu128torch2.9-cp311-cp311-linux_x86_64.whl; \
uv pip3 install --no-cache-dir flash_attn-2.8.3+cu128torch2.9-cp311-cp311-linux_x86_64.whl; \ uv pip install --no-cache-dir flash_attn-2.8.3+cu128torch2.9-cp311-cp311-linux_x86_64.whl; \
rm flash_attn-2.8.3+cu128torch2.9-cp311-cp311-linux_x86_64.whl; \ rm flash_attn-2.8.3+cu128torch2.9-cp311-cp311-linux_x86_64.whl; \
elif [ "$CUDA" = "130" ]; then \ elif [ "$CUDA" = "130" ]; then \
wget -nv https://github.com/mjun0812/flash-attention-prebuild-wheels/releases/download/v0.5.4/flash_attn-2.8.3+cu130torch2.9-cp311-cp311-linux_x86_64.whl; \ wget -nv https://github.com/mjun0812/flash-attention-prebuild-wheels/releases/download/v0.5.4/flash_attn-2.8.3+cu130torch2.9-cp311-cp311-linux_x86_64.whl; \
uv pip3 install --no-cache-dir flash_attn-2.8.3+cu130torch2.9-cp311-cp311-linux_x86_64.whl; \ uv pip install --no-cache-dir flash_attn-2.8.3+cu130torch2.9-cp311-cp311-linux_x86_64.whl; \
rm flash_attn-2.8.3+cu130torch2.9-cp311-cp311-linux_x86_64.whl; \ rm flash_attn-2.8.3+cu130torch2.9-cp311-cp311-linux_x86_64.whl; \
fi \ fi \
;; \ ;; \