Models
Datasets
Spaces
Buckets new
Docs
Enterprise
Pricing
- Website
- Community
- Solutions
Log In
Sign Up

Reza2kn
/

mega-asr-coreml

Automatic Speech Recognition

apple-neural-engine

speech-recognition

mixed-precision

Model card Files Files and versions

mega-asr-coreml

5.03 GB

Ctrl+K

Ctrl+K

1 contributor

History: 31 commits

Reza2kn's picture

Document Apache 2.0 licensing and preserve upstream notices

f67c7dd verified 24 days ago

coreml
Mixed 8/4 CoreML (90.6% on VITW): LUT-8 attn + LUT-4 MLP 4 months ago
examples
Add recording.wav 4 months ago
onnx
Add audio_encoder_fp32.onnx 4 months ago
tokenizer
Add preprocessor_config.json 4 months ago
.gitattributes

1.98 kB
Add recording.wav 4 months ago
LICENSE

11.4 kB
Document Apache 2.0 licensing and preserve upstream notices 24 days ago
README.md

8.59 kB
Document Apache 2.0 licensing and preserve upstream notices 24 days ago
convert_embeds.py

9.29 kB
Working CoreML LUT4 input_embeds variant (86.9% on VITW) 4 months ago
convert_embeds_mixed.py

9.59 kB
Mixed 8/4 CoreML (90.6% on VITW): LUT-8 attn + LUT-4 MLP 4 months ago
inference_asr.py

9.28 kB
Working CoreML LUT4 input_embeds variant (86.9% on VITW) 4 months ago