TAO example for YoloV3 in Google Colab fails to execute the training

Hello,

I’m evaluating TAO for our training needs. I’m trying to use the example for YoloV3 in the Google Colab cloud. Unfortunately, the example cannot start the training. The system complains about keras being not available. Keras seems to have been successfully installed previously so I’m a bit puzzled. Nevertheless, there are during the installation of tensorflow warnings regarding some version mismatches. I can give more details if needed. Could you give a try to check if the example is up to date with TAO 4.0 please ?

Notebook:

Step:

KITTI trainval

!tao yolo_v3 dataset_convert -d $SPECS_DIR/tfrecords_kitti_trainval.txt
-o $DATA_DIR/tfrecords/kitti_trainval/kitti_trainval

Output:
Traceback (most recent call last):
File “/usr/local/bin/yolo_v3”, line 5, in
from iva.yolo_v3.entrypoint.yolo_v3 import main
File “/usr/local/lib/python3.6/dist-packages/iva/init.py”, line 10, in
import third_party.keras.mixed_precision as MP
File “/usr/local/lib/python3.6/dist-packages/third_party/keras/mixed_precision.py”, line 11, in
import keras
ModuleNotFoundError: No module named ‘keras’

Thank you,
Regards,
Mathieu.

Well, I erased everything in my Google drive, tried again and now it’s running. Thank you.

1 Like