Add piano separation outputs (sam-audio-small, 30s clip)
#2
by fuzeey - opened
Two files produced by running facebook/sam-audio-small with 2-way tensor parallelism on a 30-second clip of elisabeth-moi-seule-choisis-ma-vie-4-juin-2022.mp3 on an AWS trn2.3xlarge:
piano_track.wav— separated piano waveform (2.8 MB, 44.1 kHz mono)piano_notes.mid— piano transcription viapiano_transcription_inference
Generated as part of Neuron TP validation in torch-neuronx.
Jingya changed pull request status to merged