Invalid decryption. Unable to open file xxx, The key used to load the model is incorrect

[jmliu_conti]

Hello,

When I followed TLT3.0 instructions for transfer learning of YOLOv4, some errors happened and I cannot continue.
When I run the following command:

tlt yolo_v4 train -e /home/ubuntu/workspace/tlt-experiments/yolo_v4/specs/yolo_v4_train_resnet18_kitti.txt -r /home/ubuntu/workspace/projects/project_egr/yolo_v4/experiment_dir_unpruned -k MY_NGC_KEY --gpus 1

Error information:

....
2021-07-30 03:12:37,573 [WARNING] tensorflow: From /usr/local/lib/python3.6/dist-packages/keras/backend/tensorflow_backend.py:2018: The name tf.image.resize_nearest_neighbor is deprecated. Please use tf.compat.v1.image.resize_nearest_neighbor instead.

Invalid decryption. Unable to open file (unable to open file: name = '/home/ubuntu/workspace/projects/project_egr/yolo_v4/pretrained_resnet18/tlt_pretrained_object_detection_vresnet18/resnet_18.hdf5', errno = 2, error message = 'No such file or directory', flags = 0, o_flags = 0). The key used to load the model is incorrect.
2021-07-30 03:12:40,097 [INFO] tlt.components.docker_handler.docker_handler: Stopping container.

I’m quite sure the path of file ‘resnet_18.hdf5’ is correct, and I have already login to nvcr.io
Could you give me some guidance how to continue, please?

Thanks.

Please provide the following information when requesting support.

• Hardware [T4 installed in AWS EC2)
• Network Type (Yolo_v4)
• TLT Version (3.0)

  Configuration of the TLT Instance
dockers: ['nvidia/tlt-streamanalytics', 'nvidia/tlt-pytorch']
format_version: 1.0
tlt_version: 3.0
published_date: 04/16/2021

• Training spec file(If have, please share here)
yolo_v4_train_resnet18_kitti.txt (2.5 KB)

• How to reproduce the issue ? (This is for errors. Please share the command line and the detailed log here.

tlt yolo_v4 train -e /home/ubuntu/workspace/tlt-experiments/yolo_v4/specs/yolo_v4_train_resnet18_kitti.txt -r /home/ubuntu/workspace/projects/project_egr/yolo_v4/experiment_dir_unpruned -k MY_NGC_KEY --gpus 1

[Morganh]

When you run tlt yolo_v4 train, the following path should be the path inside the docker instead of your local path.

Please modify your command based on the ~/.tlt_mounts.json.

[jmliu_conti]

@Morganh

e ~/.tlt_mou

Thanks for fast response.
Now I understand all the path settings are not relevant to the host, but the docker container.

Thanks

[nk_white]

I came here to post the same issue. The path setup at the beginning of the Notebooks is not obvious.

[Morganh]

Please refer to NVIDIA TAO Documentation