Sinclair Hudson comments

Results 9 comments of


                                            Sinclair Hudson

total_loss: nan?

I'm having the same issue. Could I please get an english explanation?

total_loss: nan?

@YunYang1994 could I get a quick english translation please?

total_loss: nan?

I solved the issue by reducing the learning rate and using warmup epochs. The learning rate slowly increases and then decreases, and never gets too high. This will prevent the...

AssertionError: failed to read all data

Hmmm, I'm getting the same issue. Except when I comment out the assert statement, I get the following error: > tensorflow.python.framework.errors_impl.InvalidArgumentError: Input to reshape is a tensor with 5577 values,...

AssertionError: failed to read all data

Just figured out what went wrong. In video_demo.py, you have to change the number of classess manually, on line 24. Could this be changed so that it automatically reads the...

out of memory

I am also having this issue. I can't even train b0 on a GTX 1080...

RuntimeError related to Jedi version being too old

Same issue on ubuntu 20.04... Is this a problem with Jedi or coc?

cuda device-side runtime error when training on custom dataset for JSON outputs

FYI, this dataset also doesn't work. It fails for another reason, TypeError somewhere else. This could be unrelated, but I'm wondering if the method for injecting prompts/responses is robust to...

cuda device-side runtime error when training on custom dataset for JSON outputs

[device-side.txt](https://github.com/georgian-io/LLM-Finetuning-Toolkit/files/15367611/device-side.txt) I've included the whole output, both stdout and stderr, for the "azizshaw/text_to_json" case.