Automatic Speech Recognition
NeMo
Core ML
English
speech-recognition
FastConformer
end-of-utterance
voice agent
Instructions to use FluidInference/parakeet-realtime-eou-120m-coreml with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- NeMo
How to use FluidInference/parakeet-realtime-eou-120m-coreml with NeMo:
import nemo.collections.asr as nemo_asr asr_model = nemo_asr.models.ASRModel.from_pretrained("FluidInference/parakeet-realtime-eou-120m-coreml") transcriptions = asr_model.transcribe(["file.wav"]) - Notebooks
- Google Colab
- Kaggle
Download 320ms/streaming_encoder_metadata.json from FluidInference/parakeet-realtime-eou-120m-coreml: direct link, hf CLI and curl.
- Browser
- Download file 263 Bytes
-
https://huggingface.co/FluidInference/parakeet-realtime-eou-120m-coreml/resolve/main/320ms/streaming_encoder_metadata.json
- Command line
-
hf download hf://FluidInference/parakeet-realtime-eou-120m-coreml/320ms/streaming_encoder_metadata.json
-
curl -L -o streaming_encoder_metadata.json https://huggingface.co/FluidInference/parakeet-realtime-eou-120m-coreml/resolve/main/320ms/streaming_encoder_metadata.json
263 Bytes
| { | |
| "model_id": "nvidia/parakeet_realtime_eou_120m-v1", | |
| "chunk_ms": 320, | |
| "mel_frames": 64, | |
| "pre_cache_size": 9, | |
| "valid_out_len": 4, | |
| "samples_per_chunk": 10240, | |
| "hidden_dim": 512, | |
| "num_layers": 17, | |
| "cache_channel_size": 70, | |
| "cache_time_size": 8 | |
| } |