Automatic Speech Recognition
NeMo
Safetensors
PyTorch
parakeet_tdt
speech
audio
Transducer
TDT
FastConformer
Conformer
NeMo
hf-asr-leaderboard
Transformers
Eval Results (legacy)
Eval Results
Instructions to use nvidia/parakeet-tdt-0.6b-v3 with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- NeMo
How to use nvidia/parakeet-tdt-0.6b-v3 with NeMo:
import nemo.collections.asr as nemo_asr asr_model = nemo_asr.models.ASRModel.from_pretrained("nvidia/parakeet-tdt-0.6b-v3") transcriptions = asr_model.transcribe(["file.wav"]) - Notebooks
- Google Colab
- Kaggle
GPU memory requirements
#30
by ak1020 - opened
I'm trying to run model with "NVIDIA RTX 1000 Ada Generation Laptop GPU" for inference with 6.0 GB total memory. My audio file is less than 6 minutes. I get error "RuntimeError: CUDA driver error: out of memory". If I first run the model with small example .wav and then again with my .wav, model works OK.
Use INT8 quantized version with Sherpa-onnx api
ak1020 changed discussion status to closed