Lightrun Answers was designed to reduce the constant googling that comes with debugging 3rd party libraries. It collects links to all the places you might be looking at while hunting down a tough bug.
And, if you’re still stuck at the end, we’re happy to hop on a call to see how we can help out.
Show training utterances in the training log
See original GitHub issueHi,
Some of my utterances are creating issues while training. Following is the error, it’s throwing:
RuntimeError: The size of tensor a (17243) must match the size of tensor b (2500) at non-singleton dimension 1
I want to troubleshoot, exactly which utterance is creating the problem. How can I modify the script, so, while training, it will also print utterances in the terminal, which batches are currently selected for training, or there is any other efficient way to troubleshoot the error?
Thanks
Issue Analytics
- State:
- Created 2 years ago
- Comments:6
Top Results From Across the Web
Top Related Medium Post
Top Related StackOverflow Question
Troubleshoot Live Code
Top Related Reddit Thread
Top Related Hackernoon Post
Top Related Tweet
Top Related Dev.to Post
Top Related Hashnode Post
Thanks @Gastron The issue was resolved. I use batch.id, to print the utt. There was an utt of more than 10 minutes, which is creating the issue.
Thanks a lot 😃
I think technically you should be able to run the model with a different max_length - the positional encoding is not learned. Sometimes the attention mechanism doesn’t respond well to longer sequences than seen in training.