Update tokens_per_sec to tokens_per_sec_per_gpu

[kartikayk]

Context

We log tokens_per_sec_per_gpu but this is called tokens_per_sec which causes some confusion.

Changelog

Update the name of the param being logged

Test plan

Eyes:

[pytorch-bot]

🔗 Helpful Links

🧪 See artifacts and rendered test results at hud.pytorch.org/pr/pytorch/torchtune/956

• 📄 Preview Python docs built from this PR

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

✅ No Failures

As of commit 53b194a with merge base 0217d44 ():
💚 Looks good so far! There are no failures yet. 💚

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

[RdoubleA]

would tokens_per_sec_per_gpu ever differ slightly among ranks?

[kartikayk]

would tokens_per_sec_per_gpu ever differ slightly among ranks?

it's a good question - it will, but the current nomenclature doesn't really call this difference out either. We assume rank 0 is representative (reasonable for randomly shuffled data). This just makes what we log more explicit