diff --git a/.github/workflows/test_helm.yml b/.github/workflows/test_helm.yml new file mode 100644 index 000000000..f7f4952f6 --- /dev/null +++ b/.github/workflows/test_helm.yml @@ -0,0 +1,26 @@ +name: Test HELM Integration + +on: + push: + branches: [ main ] + pull_request: + branches: [ main ] + +jobs: + compile-docs: + + runs-on: ubuntu-latest + env: + OS: ubuntu-latest + + steps: + - uses: actions/checkout@v4 + - uses: actions/setup-python@v5 + with: + python-version: '3.8' + cache: 'pip' # caching pip dependencies + - run: pip install git+https://github.com/stanford-crfm/helm.git + + - name: Test Helm + run: utils/run_helm.sh + diff --git a/.gitignore b/.gitignore index 0b9ae7722..f89426b82 100644 --- a/.gitignore +++ b/.gitignore @@ -150,3 +150,5 @@ ibmcos_datasets/ kaggle.json src/unitxt/catalog_back/* +prod_env/* +benchmark_output/* diff --git a/utils/run_helm.sh b/utils/run_helm.sh new file mode 100755 index 000000000..280d07ee5 --- /dev/null +++ b/utils/run_helm.sh @@ -0,0 +1,12 @@ +recipe="card=cards.wnli,template=templates.classification.multi_class.relation.default" +hf_model="microsoft/phi-1_5" + +helm-run \ + --run-entries "unitxt:$recipe,model=$hf_model" \ + --enable-huggingface-models $hf_model \ + --max-eval-instances 10 --suite v1 + +if [ ! -f "benchmark_output/runs/v1/unitxt:card=cards.wnli,template=templates.classification.multi_class.relation.default,model=microsoft_phi-1_5/scenario_state.json" ]; then + echo "Error: File does not exist." >&2 + exit 1 +fi \ No newline at end of file