FluidInference
/

Nemotron-3.5-ASR-Streaming-Multilingual-0.6b-ov

Automatic Speech Recognition

cache-aware-streaming

Model card Files Files and versions

Nemotron-3.5-ASR-Streaming-Multilingual-0.6b-ov / int8

784 MB

Ctrl+K

Ctrl+K

1 contributor

History: 2 commits

alexwengg's picture

Replace int8 with weight-only quantization (accuracy-preserving; full-PTQ regressed EN WER)

dd0be4f verified about 1 month ago

metadata.json

3.08 kB
Replace int8 with weight-only quantization (accuracy-preserving; full-PTQ regressed EN WER) about 1 month ago
nemotron_decoder.bin
Pickle imports
- No problematic imports detected
What is a pickle import?
29.9 MB
xet

Add INT8 (encoder) + FP16 variant about 1 month ago
nemotron_decoder.xml

28.1 kB
Add INT8 (encoder) + FP16 variant about 1 month ago
nemotron_encoder.bin
Pickle imports
- No problematic imports detected
What is a pickle import?
732 MB
xet

Replace int8 with weight-only quantization (accuracy-preserving; full-PTQ regressed EN WER) about 1 month ago
nemotron_encoder.xml

2.7 MB
Replace int8 with weight-only quantization (accuracy-preserving; full-PTQ regressed EN WER) about 1 month ago
nemotron_joint.bin
Pickle imports
- No problematic imports detected
What is a pickle import?
18.9 MB
xet

Add INT8 (encoder) + FP16 variant about 1 month ago
nemotron_joint.xml

12.9 kB
Add INT8 (encoder) + FP16 variant about 1 month ago
nemotron_preprocessor.bin
Pickle imports
- No problematic imports detected
What is a pickle import?
66.7 kB
xet

Add INT8 (encoder) + FP16 variant about 1 month ago
nemotron_preprocessor.xml

42.4 kB
Add INT8 (encoder) + FP16 variant about 1 month ago
nemotron_vocab.json

236 kB
Add INT8 (encoder) + FP16 variant about 1 month ago