Fix error reporting in TL3_EfficientDet_convergence and TL3_YOLO_conv…

…ergence (NVIDIA#3376) - due to the usage of tee the errors are not propagated correctly in TL3_EfficientDet_convergence and TL3_YOLO_convergence. The PR fixes the issue by using PIPESTATUS env variable to obtain the pipelined command result Signed-off-by: Janusz Lisiecki <jlisiecki@nvidia.com>
cyyever · Jan 23, 2022 · c213826 · c213826
1 parent 1c68bd4
commit c213826
Show file tree

Hide file tree

Showing 2 changed files with 3 additions and 3 deletions.
diff --git a/qa/TL3_EfficientDet_convergence/test_tensorflow.sh b/qa/TL3_EfficientDet_convergence/test_tensorflow.sh
@@ -26,8 +26,8 @@ python train.py
     --eval_after_training                                                                           \
     --log_dir .                                                                                     \
     --ckpt_dir .                                                                                    \
-    --output_filename out_weights_1.h5  2>&1 | tee $LOG  
+    --output_filename out_weights_1.h5  2>&1 | tee $LOG
 
 popd
 
-CLEAN_AND_EXIT 0
+CLEAN_AND_EXIT ${PIPESTATUS[0]}
diff --git a/qa/TL3_YOLO_convergence/test_tensorflow.sh b/qa/TL3_YOLO_convergence/test_tensorflow.sh
@@ -21,4 +21,4 @@ python src/main.py train \
     --eval_annotations /data/coco/coco-2017/coco2017/annotations/instances_val2017.json 2>&1 | tee $LOG
 
 
-CLEAN_AND_EXIT 0
+CLEAN_AND_EXIT ${PIPESTATUS[0]}