Feature Extraction
Transformers
Safetensors
granite_speech_nar
speech
asr
non-autoregressive
ctc
speech_recognition
automatic_speech_recognition
custom_code
Instructions to use ibm-granite/granite-speech-4.1-2b-nar with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Transformers
How to use ibm-granite/granite-speech-4.1-2b-nar with Transformers:
# Use a pipeline as a high-level helper from transformers import pipeline pipe = pipeline("feature-extraction", model="ibm-granite/granite-speech-4.1-2b-nar", trust_remote_code=True)# Load model directly from transformers import AutoModel model = AutoModel.from_pretrained("ibm-granite/granite-speech-4.1-2b-nar", trust_remote_code=True, dtype="auto") - Notebooks
- Google Colab
- Kaggle
Add native transformers GraniteSpeechNarForCTC format
#6
by eustlb HF Staff - opened
Converts this checkpoint from the remote-code implementation to the native GraniteSpeechNarForCTC model now available in π€ Transformers.
Changes
- Native
config.json(model_type: granite_speech_nar, noauto_map) and re-keyedmodel.safetensors. - Processor/feature-extractor config in
processor_config.json(nativeSequenceFeatureExtractorlayout). - Remove the remote-code modules (
*_granite_speech_nar.py,__init__.py) and the stalepreprocessor_config.json(superseded by the nested feature-extractor config).
Produced by transformers/src/transformers/models/granite_speech_nar/convert_granite_speech_nar_to_hf.py.
Follow-ups for maintainers: model.sig needs re-signing (weights changed), and the README usage snippet can drop trust_remote_code.