Lightning-AI / Lightning-AI/lit-llama

[question] nan loss value and run time error

Open
#463 0 comments 0 reactions 0 assignees View on GitHub

Nobody has claimed this yet.

Dominant language
Python
Stars
6.1k
Forks
517
PR merge metrics
No merged PRs in 30d

Description

Dear all,

I finetuning mydata with open llama. While running finetune/lora.py, I don't see the loss number as follows:
...
iter 3198: loss nan, time: 134.94ms

and while validating, it ends up with an error:
...
File ".../lit-llama/generate.py", line 74, in generate
idx_next = torch.multinomial(probs, num_samples=1).to(dtype=dtype)
RuntimeError: probability tensor contains either `inf`, `nan` or element < 0

Could you tell me how I can resolve this?

Thanks in advance.

Contributor guide

No contributing guide indexed for this repository

First steps

  1. Read the whole issue, then the project's contributing guide.
  2. Comment on the issue to say you are picking it up — it saves two people doing the same work.
  3. Fork the repository and make your change on a branch.
  4. Open a pull request that references the issue number.

Research direction

Start by reproducing the reported NaN loss while running finetune/lora.py, then inspect generate.py at line 74 where torch.multinomial fails. Determine the input conditions that produce invalid probabilities during validation; done means the cause is isolated and a verified resolution is documented.

Written by the indexing model from the issue text.

Assessment

Tech stack
python, pytorch
Domain
machine-learning
Issue type
Bug
Difficulty
4/5
Estimated time
3-5 days
Activity status
Stale
Clarity
Needs clarification
Newbie friendliness
20/100

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.