Granite speech speedup + model saving bugfix #39028
ensure the query is updated during training
db4a4af8
avoid a crash when `kwargs` contain `padding=True`
8ee3429d
minor
8dec2bae
Remove mel_spec lazy init, and rename to mel_filters.
4db4c994
minor - most feature extractors has a `sampling_rate` property
98844ec5
Merge branch 'main' into granite_speech_updates
6e68d8ca
speedup relative position embeddings
7064db7c
Merge branch 'huggingface:main' into granite_speech_updates
e94d0a18
fix several issues in model saving/loading:
9c06f95a
Merge branch 'granite_speech_updates' of https://github.com/avihu111/…
6c2db624
minor
8b79d9e7
minor
313e4a2e
minor
29a69be7
avihu111
marked this pull request as draft 189 days ago
fixing a crash without peft active
bc152b9b
avihu111
changed the title Granite speech updates Granite speech speedup + model saving bugfix 188 days ago
avihu111
marked this pull request as ready for review 188 days ago
add todo to replace einsum
0ebb8f07
Merge branch 'main' into granite_speech_updates
04911237
Assignees
No one assigned
Login to write a write a comment.
Login via GitHub