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

Add ci and images for CUDA 12.8 for B200s (#2683) [skip ci]

Browse Source 
* Add ci and images for CUDA 12.8 for B200s

* add comments explaining CI [skip e2e]
This commit is contained in:
Wing Lian
2025-05-16 13:06:08 -04:00
committed by GitHub
parent 86472715da
commit 8f8a7afb05
2 changed files with 19 additions and 0 deletions
Download Patch File Download Diff File Expand all files Collapse all files

10
.github/workflows/main.yml vendored
View File

@@ -31,6 +31,11 @@ jobs:
python_version: "3.11" python_version: "3.11"
pytorch: 2.7.0 pytorch: 2.7.0
axolotl_extras: axolotl_extras:
- cuda: 128
cuda_version: 12.8.1
python_version: "3.11"
pytorch: 2.7.0
axolotl_extras:
runs-on: axolotl-gpu-runner runs-on: axolotl-gpu-runner
steps: steps:
- name: Checkout - name: Checkout
@@ -94,6 +99,11 @@ jobs:
python_version: "3.11" python_version: "3.11"
pytorch: 2.7.0 pytorch: 2.7.0
axolotl_extras: axolotl_extras:
- cuda: 128
cuda_version: 12.8.1
python_version: "3.11"
pytorch: 2.7.0
axolotl_extras:
runs-on: axolotl-gpu-runner runs-on: axolotl-gpu-runner
steps: steps:
- name: Checkout - name: Checkout

9
.github/workflows/tests.yml vendored
View File

@@ -295,6 +295,7 @@ jobs:
find "$(pip cache dir)/http-v2" -type f -mtime +14 -exec rm {} \; find "$(pip cache dir)/http-v2" -type f -mtime +14 -exec rm {} \;
docker-e2e-tests-1st: docker-e2e-tests-1st:
# Run this job first as a gate for running the remainder of the test matrix
if: ${{ ! contains(github.event.commits[0].message, '[skip e2e]') && github.repository_owner == 'axolotl-ai-cloud' }} if: ${{ ! contains(github.event.commits[0].message, '[skip e2e]') && github.repository_owner == 'axolotl-ai-cloud' }}
# this job needs to be run on self-hosted GPU runners... # this job needs to be run on self-hosted GPU runners...
runs-on: [self-hosted, modal] runs-on: [self-hosted, modal]
@@ -341,6 +342,8 @@ jobs:
# this job needs to be run on self-hosted GPU runners... # this job needs to be run on self-hosted GPU runners...
runs-on: [self-hosted, modal] runs-on: [self-hosted, modal]
timeout-minutes: 90 timeout-minutes: 90
# Only run the remainder of the matrix if the first e2e check passed;
# this is to save on wasted compute costs for known failures that get caught in the first run
needs: [pre-commit, pytest, docker-e2e-tests-1st] needs: [pre-commit, pytest, docker-e2e-tests-1st]
strategy: strategy:
@@ -365,6 +368,12 @@ jobs:
pytorch: 2.7.0 pytorch: 2.7.0
num_gpus: 1 num_gpus: 1
axolotl_extras: axolotl_extras:
- cuda: 128
cuda_version: 12.8.1
python_version: "3.11"
pytorch: 2.7.0
num_gpus: 1
axolotl_extras:
steps: steps:
- name: Checkout - name: Checkout
uses: actions/checkout@v4 uses: actions/checkout@v4