Enable compile on MPS with experimental warning

[manuelcandales]

This enables compile on MPS, modifying the warning message to say this is experimental and is not ready for broad use yet
We also adjust inductor's config to avoid the inductor's decomposition of mm if running on MPS backend

python3 torchchat.py generate stories15M --dtype float16 --prompt "Once upon a time" --num-samples 3 --device mps --compile

[pytorch-bot]

🔗 Helpful Links

🧪 See artifacts and rendered test results at hud.pytorch.org/pr/pytorch/torchchat/1523

• 📄 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 d491b34 with merge base 1384f7d ():
💚 Looks good so far! There are no failures yet. 💚

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

[Jack-Khuu]

We added this for MPS LLama 3.2-Vision. Does this affect your work here?

torchchat/torchchat.py

Line 15 in 2640f6a

os.environ["PYTORCH_ENABLE_MPS_FALLBACK"] = "1"

[manuelcandales]

PYTORCH_ENABLE_MPS_FALLBACK

I don't think so. Will experiment with it, but shouldn't have any effect on stories or text only llamas