Automatic Speech Recognition
NeMo
Safetensors
PyTorch
fastconformer
automatic-speech-translation
speech
audio
Transformer
FastConformer
Conformer
NeMo
hf-asr-leaderboard
Eval Results (legacy)
Eval Results
Instructions to use nvidia/canary-1b-flash with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- NeMo
How to use nvidia/canary-1b-flash with NeMo:
import nemo.collections.asr as nemo_asr asr_model = nemo_asr.models.ASRModel.from_pretrained("nvidia/canary-1b-flash") transcriptions = asr_model.transcribe(["file.wav"]) - Notebooks
- Google Colab
- Kaggle
| { | |
| "activation_dropout": 0.1, | |
| "activation_function": "silu", | |
| "architectures": [ | |
| "FastConformerModel" | |
| ], | |
| "attention_dropout": 0.1, | |
| "bos_token_id": 1, | |
| "conv_kernel_size": 9, | |
| "d_model": 1024, | |
| "dropout": 0.1, | |
| "dropout_emb": 0.0, | |
| "encoder_attention_heads": 8, | |
| "encoder_ffn_dim": 4096, | |
| "encoder_layerdrop": 0.1, | |
| "encoder_layers": 32, | |
| "eos_token_id": 2, | |
| "initializer_range": 0.02, | |
| "model_type": "fastconformer", | |
| "nemo_decoder_type": "none", | |
| "nemo_encoder_type": "nemo.collections.asr.modules.ConformerEncoder", | |
| "nemo_model_name": "nvidia/canary-1b-flash", | |
| "nemo_model_type": "canary", | |
| "num_mel_bins": 128, | |
| "pad_token_id": 0, | |
| "subsampling_conv_channels": 256, | |
| "subsampling_factor": 8, | |
| "torch_dtype": "float32", | |
| "transformers_version": "4.53.0.dev0", | |
| "use_bias": true, | |
| "vocab_size": 1024, | |
| "xscaling": false | |
| } | |