use stable lm to test pr
This commit is contained in:
parent
e08c74f1d1
commit
8472de90e8
1 changed files with 7 additions and 7 deletions
14
.github/workflows/llm-harness-evaluation.yml
vendored
14
.github/workflows/llm-harness-evaluation.yml
vendored
|
|
@ -68,9 +68,9 @@ jobs:
|
||||||
- name: set-pr-env
|
- name: set-pr-env
|
||||||
if: ${{github.event_name == 'pull_request'}}
|
if: ${{github.event_name == 'pull_request'}}
|
||||||
env:
|
env:
|
||||||
PR_MATRIX_MODEL_NAME: '["Mistral-7B-v0.1"]'
|
PR_MATRIX_MODEL_NAME: '["stablelm-3b-4e1t"]'
|
||||||
PR_MATRIX_TASK: '["arc", "truthfulqa", "winogrande"]'
|
PR_MATRIX_TASK: '["truthfulqa", "winogrande"]'
|
||||||
PR_MATRIX_PRECISION: '["fp8"]'
|
PR_MATRIX_PRECISION: '["sym_int4"]'
|
||||||
PR_LABELS: '["self-hosted", "llm", "temp-arc01"]'
|
PR_LABELS: '["self-hosted", "llm", "temp-arc01"]'
|
||||||
|
|
||||||
run: |
|
run: |
|
||||||
|
|
@ -294,7 +294,7 @@ jobs:
|
||||||
- name: Download fp16.csv for summary
|
- name: Download fp16.csv for summary
|
||||||
shell: bash
|
shell: bash
|
||||||
run: |
|
run: |
|
||||||
wget https://raw.githubusercontent.com/intel-analytics/BigDL/main/python/llm/test/benchmark/harness/fp16.csv -O ${{ env.NIGHTLY_FOLDER}}/../fp16.csv
|
wget https://raw.githubusercontent.com/intel-analytics/BigDL/main/python/llm/dev/benchmark/harness/fp16.csv -O ${{ env.NIGHTLY_FOLDER}}/../fp16.csv
|
||||||
ls ${{ env.NIGHTLY_FOLDER}}/..
|
ls ${{ env.NIGHTLY_FOLDER}}/..
|
||||||
|
|
||||||
- name: Summarize the results for nightly run
|
- name: Summarize the results for nightly run
|
||||||
|
|
@ -304,8 +304,8 @@ jobs:
|
||||||
ls /home/arda/harness-action-runners/nightly-accuracy-data/${{ env.OUTPUT_PATH }}
|
ls /home/arda/harness-action-runners/nightly-accuracy-data/${{ env.OUTPUT_PATH }}
|
||||||
pip install pandas==1.5.3
|
pip install pandas==1.5.3
|
||||||
python ${{ github.workspace }}/python/llm/dev/benchmark/harness/make_table_and_csv.py ${{ env.NIGHTLY_FOLDER}}/${{ env.OUTPUT_PATH }} ${{ env.NIGHTLY_FOLDER}}
|
python ${{ github.workspace }}/python/llm/dev/benchmark/harness/make_table_and_csv.py ${{ env.NIGHTLY_FOLDER}}/${{ env.OUTPUT_PATH }} ${{ env.NIGHTLY_FOLDER}}
|
||||||
python ${{ github.workspace }}/python/llm/test/benchmark/harness/harness_csv_to_html.py -f ${{ env.NIGHTLY_FOLDER}}
|
python ${{ github.workspace }}/python/llm/dev/benchmark/harness/harness_csv_to_html.py -f ${{ env.NIGHTLY_FOLDER}}
|
||||||
python ${{ github.workspace }}/python/llm/test/benchmark/harness/update_html_in_parent_folder.py -f ${{ env.NIGHTLY_FOLDER }}
|
python ${{ github.workspace }}/python/llm/dev/benchmark/harness/update_html_in_parent_folder.py -f ${{ env.NIGHTLY_FOLDER }}
|
||||||
|
|
||||||
- name: Summarize the results for pull request
|
- name: Summarize the results for pull request
|
||||||
if: github.event_name == 'pull_request'
|
if: github.event_name == 'pull_request'
|
||||||
|
|
@ -314,4 +314,4 @@ jobs:
|
||||||
ls /home/arda/harness-action-runners/pr-accuracy-data/${{ env.OUTPUT_PATH }}
|
ls /home/arda/harness-action-runners/pr-accuracy-data/${{ env.OUTPUT_PATH }}
|
||||||
pip install pandas==1.5.3
|
pip install pandas==1.5.3
|
||||||
python ${{ github.workspace }}/python/llm/dev/benchmark/harness/make_table_and_csv.py ${{ env.PR_FOLDER}}/${{ env.OUTPUT_PATH }} ${{ env.PR_FOLDER}}
|
python ${{ github.workspace }}/python/llm/dev/benchmark/harness/make_table_and_csv.py ${{ env.PR_FOLDER}}/${{ env.OUTPUT_PATH }} ${{ env.PR_FOLDER}}
|
||||||
python ${{ github.workspace }}/python/llm/test/benchmark/harness/harness_csv_to_html.py -f ${{ env.PR_FOLDER}}
|
python ${{ github.workspace }}/python/llm/dev/benchmark/harness/harness_csv_to_html.py -f ${{ env.PR_FOLDER}}
|
||||||
Loading…
Reference in a new issue