Deep learning model debugging

Question

I am using the kaggles environment to train my model using pytorch and mltu. While the training is going on after 3 hours it shows sad emoji and still running in the background when I checked the log file it was stating error to the model2onnx is giving trouble of symbolic error. So I removed it and it still gives sad emoji. Can any one please help resolve this issue. I'm ready to share my code for debugging.