fix @olupton comments

Author: Steboss

.github/container/test-axlearn.sh

@@ -219,18 +219,22 @@ done
 # SUMMARY STATUS
 passed=0
 failed=0
+error=0
 skipped=0
 for log in ${LOG_DIRECTORY}/log_*.log; do
     count_pass=$(grep -Eo '[0-9]+ passed' "${log}" | awk '{print $1}' || true)
     count_fail=$(grep -Eo '[0-9]+ failed' "${log}" | awk '{print $1}' || true)
+    count_error=$(grep -Eo '[0-9]+ error' "${log}" | awk '{print $1}' || true)
     count_skipped=$(grep -Eo '[0-9]+ skipped' "${log}" | awk '{print $1}' || true)
     # in case of None
     count_pass=${count_pass:-0}
     count_fail=${count_fail:-0}
+    count_error=${count_error:-0}
     count_skipped=${count_skipped:-0}
     # count all the tests
     (( passed += count_pass ))
     (( failed += count_fail ))
+    (( failed += count_error ))
     (( skipped += count_skipped ))
 done
 

.github/eks-workflow-files/axlearn/axlearn-job.yml

@@ -22,7 +22,7 @@ spec:
                       LOG_DIR="/output/${RUN_ID}"
                       mkdir -p ${LOG_DIR}
                       # test on JAX, make sure 8 devices are visible
-                      pytest-xdist.sh 8 8 ${LOG_DIR}/axlearn-unittests.jsonl test-axlearn.sh --directory "." --output ${LOG_DIR} --test-files "/opt/axlearn/axlearn/common/*_test.py"
+                      pytest-xdist.sh 8 4 ${LOG_DIR}/axlearn-unittests.jsonl test-axlearn.sh --directory "." --output ${LOG_DIR} --test-files "/opt/axlearn/axlearn/common/*_test.py"
                   env:
                     - name: RUN_ID
                       value: PLACEHOLDER

.github/workflows/_ci.yaml

@@ -616,14 +616,14 @@ jobs:
       id: log-s3
       run: |
         mkdir -p axlearn-output
-        aws s3 cp s3://jax-toolbox-eks-output/axlearn/${{ github.run_id }}/logs/summary.txt axlearn_output/
-        aws s3 cp s3://jax-toolbox-eks-output/axlearn/${{ github.run_id }}/logs/ axlearn_output/ --recursive --exclude "*" --include "*.log"
-        aws s3 cp s3://jax-toolbox-eks-output/axlearn/${{ github.run_id }}/logs/ axlearn_output/ --recursive --exclude "*" --include "*.xml"
+        aws s3 cp s3://jax-toolbox-eks-output/axlearn/${{ github.run_id }}/logs/summary.txt axlearn-output/
+        aws s3 cp s3://jax-toolbox-eks-output/axlearn/${{ github.run_id }}/logs/ axlearn-output/ --recursive --exclude "*" --include "*.log"
+        aws s3 cp s3://jax-toolbox-eks-output/axlearn/${{ github.run_id }}/logs/ axlearn-output/ --recursive --exclude "*" --include "*.xml"
 
 
-        passed_tests=$(grep -Eo 'PASSED:[[:space:]]*[0-9]+' axlearn_output/summary.txt | grep -Eo '[0-9]+' )
-        failed_tests=$(grep -Eo 'FAILED:[[:space:]]*[0-9]+' axlearn_output/summary.txt | grep -Eo '[0-9]+' )
-        skipped_tests=$(grep -Eo 'SKIPPED:[[:space:]]*[0-9]+' axlearn_output/summary.txt | grep -Eo '[0-9]+' )
+        passed_tests=$(grep -Eo 'PASSED:[[:space:]]*[0-9]+' axlearn-output/summary.txt | grep -Eo '[0-9]+' )
+        failed_tests=$(grep -Eo 'FAILED:[[:space:]]*[0-9]+' axlearn-output/summary.txt | grep -Eo '[0-9]+' )
+        skipped_tests=$(grep -Eo 'SKIPPED:[[:space:]]*[0-9]+' axlearn-output/summary.txt | grep -Eo '[0-9]+' )
         total_tests=$((failed_tests + passed_tests + skipped_tests))
 
         echo "Passed tests: $passed_tests"