File size: 2,356 Bytes
98a6067 e8ee9a7 98a6067 e8ee9a7 98a6067 e8ee9a7 98a6067 |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 |
import gradio as gr
import numpy as np
import onnx_asr
models = {
name: onnx_asr.load_model(name)
for name in [
"gigaam-v2-ctc",
"gigaam-v2-rnnt",
"nemo-fastconformer-ru-ctc",
"nemo-fastconformer-ru-rnnt",
"alphacep/vosk-model-ru",
"alphacep/vosk-model-small-ru",
"whisper-base",
]
}
def recoginize(audio: tuple[int, np.ndarray]):
sample_rate, waveform = audio
waveform = waveform.astype(np.float32) / 2 ** (8 * waveform.itemsize - 1)
return [[name, model.recognize(waveform, sample_rate=sample_rate, language="ru")] for name, model in models.items()]
demo = gr.Interface(
fn=recoginize,
title="ASR demo using onnx-asr (Russian models)",
description="""# Automatic Speech Recognition in Python using ONNX models - [onnx-asr](https://github.com/istupakov/onnx-asr)
## Models used in demo:
* `gigaam-v2-ctc` - Sber GigaAM v2 CTC ([origin](https://github.com/salute-developers/GigaAM), [onnx](https://huggingface.co/istupakov/gigaam-v2-onnx))
* `gigaam-v2-rnnt` - Sber GigaAM v2 RNN-T ([origin](https://github.com/salute-developers/GigaAM), [onnx](https://huggingface.co/istupakov/gigaam-v2-onnx))
* `nemo-fastconformer-ru-ctc` - Nvidia FastConformer-Hybrid Large (ru) with CTC decoder ([origin](https://huggingface.co/nvidia/stt_ru_fastconformer_hybrid_large_pc), [onnx](https://huggingface.co/istupakov/stt_ru_fastconformer_hybrid_large_pc_onnx))
* `nemo-fastconformer-ru-rnnt` - Nvidia FastConformer-Hybrid Large (ru) with RNN-T decoder ([origin](https://huggingface.co/nvidia/stt_ru_fastconformer_hybrid_large_pc), [onnx](https://huggingface.co/istupakov/stt_ru_fastconformer_hybrid_large_pc_onnx))
* `alphacep/vosk-model-ru` - Alpha Cephei Vosk 0.54-ru ([origin](https://huggingface.co/alphacep/vosk-model-ru))
* `alphacep/vosk-model-small-ru` - Alpha Cephei Vosk 0.52-small-ru ([origin](https://huggingface.co/alphacep/vosk-model-small-ru))
* `whisper-base` - OpenAI Whisper Base exported with onnxruntime ([origin](https://huggingface.co/openai/whisper-base), [onnx](https://huggingface.co/istupakov/whisper-base-onnx))
""",
inputs=[gr.Audio(min_length=1, max_length=10)],
outputs=[gr.Dataframe(headers=["Model", "result"], wrap=True, show_fullscreen_button=True)],
flagging_mode="never",
)
demo.launch()
|