update torchao pin: optimized mps lowbit shaders

[manuelcandales]

Updates torchao pin to benefit from optimizations to the MPS experimental lowbit kernels AO PR #1422

Llama 3.2 1B (llama3.2-1b-base):
1-bit: 28.0688
2-bit: 31.2422
3-bit: 30.1294
4-bit: 30.7905
5-bit: 28.1504
6-bit: 28.4321
7-bit: 27.3991

Llama 3.1 8B (llama3.1-base):
1-bit: 7.4459
2-bit: 15.6508
3-bit: 15.3086
4-bit: 16.1268
5-bit: 6.7308
6-bit: 6.4887
7-bit: 6.4537

[pytorch-bot]

🔗 Helpful Links

🧪 See artifacts and rendered test results at hud.pytorch.org/pr/pytorch/torchchat/1428

• 📄 Preview Python docs built from this PR

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

❌ 2 New Failures, 1 Pending, 2 Unrelated Failures

As of commit 3c0e898 with merge base 56be609 ():

NEW FAILURES - The following jobs have failed:

• pull / test-gpu-aoti-bfloat16 (cuda, stories15M) / linux-job (gh)
  RuntimeError: Command docker exec -t 1df37accfa2a539b4db1dfacbd7e3f9ccdb37535748ea2979a544e1adde144af /exec failed with exit code 1
• Run the aoti runner with CUDA using stories / test-runner-aot-cuda / linux-job (gh)
  RuntimeError: Command docker exec -t 298d104bbc0bf1434b8be779cc0b6ee477c9db56693875edecb8459636a2e257 /exec failed with exit code 134

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

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

• pull / test-gpu-aoti-float16 (cuda, stories15M) / linux-job (gh) (trunk failure)
  RuntimeError: run_func_( container_handle_, input_handles.data(), input_handles.size(), output_handles.data(), output_handles.size(), reinterpret_cast<AOTInductorStreamHandle>(stream_handle), proxy_executor_handle_) API call failed at /pytorch/torch/csrc/inductor/aoti_runner/model_container_runner.cpp, line 107
• pull / test-gpu-aoti-float32 (cuda, stories15M) / linux-job (gh) (trunk failure)
  RuntimeError: run_func_( container_handle_, input_handles.data(), input_handles.size(), output_handles.data(), output_handles.size(), reinterpret_cast<AOTInductorStreamHandle>(stream_handle), proxy_executor_handle_) API call failed at /pytorch/torch/csrc/inductor/aoti_runner/model_container_runner.cpp, line 107

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

[Jack-Khuu]

Lgtm, if no objections from quant folk