When I am train the yolov7 using train.py this error came could you help to solve the issue

[chintha-3]

Traceback (most recent call last):
File "train.py", line 616, in
train(hyp, opt, device, tb_writer)
File "train.py", line 363, in train
loss, loss_items = compute_loss_ota(pred, targets.to(device), imgs) # loss scaled by batch_size
File "/content/drive/MyDrive/Final year project/yolov7/utils/loss.py", line 585, in call
bs, as_, gjs, gis, targets, anchors = self.build_targets(p, targets, imgs)
File "/content/drive/MyDrive/Final year project/yolov7/utils/loss.py", line 753, in build_targets
_, cost_argmin = torch.min(cost[:, anchor_matching_gt > 1], dim=0)
RuntimeError: CUDA error: device-side assert triggered

[SkalskiP]

Hi, @chintha-3! 👋🏻 Thanks for moving that question from YouTube to GitHub. Is that error in our official notebook, or is that your custom code?

[chintha-3]

Please give solution brother

[SkalskiP]

Can you zip the whole runs directory. And upload it here? I’d need to take a look at trading data to tell more.

[chintha-3]

runs zip file is very high how to upload it

[chintha-3]

Can you zip the whole runs directory. And upload it here? I’d need to take a look at trading data to tell more.

Can you suggest how to upload runs directory

[SkalskiP]

Hi 👋🏻 ! I was supper busy. Yo can:

1. zip it and upload it to your drive, and send me a link here
2. zip everything other than weights - those files weight the most and at the same time, are the least interesting to me