Include audio preprocessing for raw audio tensor

[jackzhxng]

Summary

Runs audio preprocessing (mel spectrogram conversion) on raw audio tensor .bin file, using an exported .pte from https://github.com/pytorch/executorch/blob/main/extension/audio/mel_spectrogram.py

Current limitations - no batching of output in the spectrogram processing module, so can only support audio of <30 seconds.

The speaker in this audio seems to be talking about their concerns about a device called the model or maybe they're just talking about the model in general. They mention that the model was trained with the speaker for inference, which suggests that
 the model was trained based on the speaker's data or instructions. They also mention that the volume is quite small, which could imply that the speaker is trying to control the volume of the model's output, likely because they are concerned about how loud the model's responses might
PyTorchObserver {"prompt_tokens":388,"generated_tokens":99,"model_load_start_ms":0,"model_load_end_ms":0,"inference_start_ms":1756351346381,"inference_end_ms":1756351362602,"prompt_eval_end_ms":1756351351435,"first_token_ms":1756351351435,"aggregate_sampling_time_ms":99,"SCALING_FACTOR_UNITS_PER_SECOND":1000}
I 00:00:24.036773 executorch:stats.h:104]       Prompt Tokens: 388    Generated Tokens: 99
I 00:00:24.036800 executorch:stats.h:110]       Model Load Time:                0.000000 (seconds)
I 00:00:24.036805 executorch:stats.h:117]       Total inference time:           16.221000 (seconds)              Rate:  6.103200 (tokens/second)
I 00:00:24.036815 executorch:stats.h:127]               Prompt evaluation:      5.054000 (seconds)               Rate:  76.770875 (tokens/second)
I 00:00:24.036819 executorch:stats.h:136]               Generated 99 tokens:    11.167000 (seconds)              Rate:  8.865407 (tokens/second)
I 00:00:24.036822 executorch:stats.h:147]       Time to first generated token:  5.054000 (seconds)
I 00:00:24.036828 executorch:stats.h:153]       Sampling time over 487 tokens:  0.099000 (seconds)

Stack from ghstack (oldest at bottom):

• -> Include audio preprocessing for raw audio tensor #13752
• Make token and stat callback optional in multimodal runner #13664
• Add Voxtral runner #13663
• Add audio to multimodal runner #13662

[pytorch-bot]

🔗 Helpful Links

🧪 See artifacts and rendered test results at hud.pytorch.org/pr/pytorch/executorch/13752

• 📄 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

As of commit ca19f68 with merge base 99e6349 ():

NEW FAILURE - The following job has failed:

• Build documentation / build (buck2) / Build doc (gh)
  At least one of the pre-conditions you specified did not hold

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