EfficientNet Error File "/app/train.py

Report bugs
#1

2025-11-25 13:52:26.265169: W tensorflow/compiler/tf2tensorrt/utils/py_utils.cc:38] TF-TRT Warning: Cannot dlopen some TensorRT libraries. If you would like to use Nvidia GPU with TensorRT, please make sure the missing libraries mentioned above are installed properly.
2025-11-25 13:52:27.373249: E tensorflow/compiler/xla/stream_executor/cuda/cuda_driver.cc:267] failed call to cuInit: UNKNOWN ERROR (34)
Training info:
Model: B0 (using pretrained weights, freezing bottom 90% of layers)
Epochs: 20
Learning rate: 0.001
Early stopping: yes (patience: 5, min delta: 0.001)
Last layers: dense: 32, dropout: 0.1
Data augmentation: None
Batch size: 32
Training on: cpu

Traceback (most recent call last):
File “/app/train.py”, line 119, in
enable_tensorboard=input.tensorboardLogging)
AttributeError: ‘types.SimpleNamespace’ object has no attribute ‘tensorboardLogging’
Application exited with code 1
Application exited with code 1

Help-me Please!!

#2

Hi,

We realised this earlier today. Will be getting a fix out in the next hour or so. Will update this thread.

Cheers,

#3

Hi @Mscfalcao I’ve pushed a new version of the EfficientNet block, this fixes your issue.

Other blocks (e.g. custom blocks that people pushed themselves) might have the same issue - a proper fix was merged in this morning, but it’ll take a few more hours to be deployed.

2 Likes