Update yaml configs. Add validation dataset

[krammnic]

Context

What is the purpose of this PR? Is it to

• add a new feature
• fix a bug
• update tests and/or documentation
• other (please add here)

Please link to any issues this PR addresses.

Changelog

What are the changes made in this PR?

• Add validation dataset loss to distributed SFT recipies #2464

Test plan

Please make sure to do each of the following if applicable to your PR. If you're unsure about any one of these just ask and we will happily help. We also have a contributing page for some guidance on contributing.

• run pre-commit hooks and linters (make sure you've first installed via pre-commit install)
• add unit tests for any new functionality
• update docstrings for any new or updated methods or classes
• run unit tests via pytest tests
• run recipe tests via pytest tests -m integration_test
• manually run any new or modified recipes with sufficient proof of correctness
• include relevant commands and any other artifacts in this summary (pastes of loss curves, eval results, etc.)

UX

If your function changed a public API, please add a dummy example of what the user experience will look like when calling it.
Here is a docstring example
and a tutorial example

• I did not change any public API
• I have added an example to docs or docstrings

[pytorch-bot]

🔗 Helpful Links

🧪 See artifacts and rendered test results at hud.pytorch.org/pr/pytorch/torchtune/2608

• 📄 Preview Python docs built from this PR

Note: Links to docs will display an error until the docs builds have been completed.

❌ 1 New Failure, 1 Cancelled Job, 1 Unrelated Failure

As of commit 154ab9d with merge base 529d2d8 ():

NEW FAILURE - The following job has failed:

• GPU tests / gpu_test (3.11, stable) (gh)
  tests/recipes/test_full_finetune_distributed.py::TestFullFinetuneDistributedRecipe::test_loss_2d_parallel[llama3/8B_full-llama3-tune-4-1-True-2]

CANCELLED JOB - The following job was cancelled. Please retry:

• GPU tests / gpu_test (3.10, stable) (gh)
  tests/recipes/test_full_finetune_distributed.py::TestFullFinetuneDistributedRecipe::test_loss_2d_parallel[llama3/8B_full-llama3-tune-4-1-True-2]

BROKEN TRUNK - The following job failed but were present on the merge base:

👉 Rebase onto the `viable/strict` branch to avoid these failures

• GPU tests / gpu_test (3.9, stable) (gh) (trunk failure)
  ##[error]The operation was canceled.

This comment was automatically generated by Dr. CI and updates every 15 minutes.

[bzz]

this looks like an accidental change and probably could be worth putting back for the PPO recipe to work

[krammnic]

Good catch!

[SalmanMohammadi]

For many of the larger models this batch size may be too high and result in OOM.

[SalmanMohammadi]

@krammnic thank you for opening this PR. One last change: could you actually set batch_size_val: ${batch_size} in every config? That way we minimise the risk of OOMing during validation, and users may increase it if they they wish.

I'm happy to merge after this.

[krammnic]

@krammnic thank you for opening this PR. One last change: could you actually set batch_size_val: ${batch_size} in every config? That way we minimise the risk of OOMing during validation, and users make increase it if they they wish.

I'm happy to merge after this.

Hey, sure!

[krammnic]

@SalmanMohammadi This is really weird

[SalmanMohammadi]

I think you also need to remove the fields from configs for recipes which we don't support validation datasets for: QAT, and LoRA and FFT single device.