search

sovit-123 / fasterrcnn-pytorch-training-pipeline

PyTorch Faster R-CNN Object Detection on Custom Dataset
MIT License
223 stars 77 forks source link

Your latest change has spoilt custom training #70

Closed Engineer-D closed 1 year ago

Engineer-D commented 1 year ago

Traceback (most recent call last): File "/content/fastercnn-pytorch-training-pipeline/train.py", line 542, in main(args) File "/content/fastercnn-pytorch-training-pipeline/train.py", line 494, in main wandb_log( File "/content/fastercnn-pytorch-training-pipeline/utils/logging.py", line 185, in wandb_log if len(val_pred_image) == 1: TypeError: object of type 'NoneType' has no len() wandb: Waiting for W&B process to finish... (failed 1).

this is what I keep getting, I ran it yesterday and it worked today it isn't

Engineer-D commented 1 year ago

I mean, I reverted to the previous code and it's training now. So something changed in engine.py has definitely affected what's going on in training.py and logging.py. logging.py is where the error is coming from.

sovit-123 commented 1 year ago

Got it, I think I know where the issue lies. I will confirm back here once the issue is fixed.

sovit-123 commented 1 year ago

@Engineer-D It's corrected. You may pull the new code now. And please close the issues if it solves the problem.