Ian Sampson

icon indicating an email [email protected]

icon location Victoria, BC icon location ML researcher and independent developer. Creator of Hush (hushaudioapp.com). Writer/scholar with a PhD from Brown University.

Results 4 comments of Ian Sampson

PyTorch convert function for op 'pad' not implemented.

I’ll second the request for an official `pad` implementation when PyTorch 1.12.0 support rolls out, though as a workaround I’ve been able to compile `torchaudio.models.Conformer` to CoreML with PyTorch 1.11.0...

PyTorch convert function for op 'pad' not implemented.

> @iansampson thanks for the info. I'm trying to convert another implementation from nvidia/NeMo instead. > Were you also able to run it/any console warnings? > Are you using it...

Converted PyTorch model is significantly larger and slower than the original

Thanks for the quick response! Just tested with the latest release (6.0b1) and yes, the results are pretty much the same (prediction time: 106.645615 s; mlpackage size: 538.8 MB).

Converted PyTorch model is significantly larger and slower than the original

I think I’ve traced the problem to the GRU layer used in TransformerEncoderLayer (aia_inter_new.py, line 41). Here’s a simple script that reproduces the problem: ``` python import time import numpy...