File size: 1,434 Bytes
7b0dc9d 104123a 7b0dc9d b51dfa4 a9a94f3 b51dfa4 b7d0d3d b51dfa4 104123a b51dfa4 b7d0d3d 104123a b51dfa4 |
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 torch
import gradio as gr
from transformers import pipeline
from scipy.io import wavfile
MODEL_NAME = "openai/whisper-large-v3"
BATCH_SIZE = 8
device = 0 if torch.cuda.is_available() else "cpu"
pipe = pipeline(
task="automatic-speech-recognition",
model=MODEL_NAME,
chunk_length_s=30,
device=device,
)
def transcribe_simple(inputs_path, task):
if inputs_path is None:
raise gr.Error("No audio file submitted! Please upload or record an audio file before submitting your request.")
sampling_rate, inputs = wavfile.read(inputs_path)
out = pipe(inputs_path, batch_size=BATCH_SIZE, generate_kwargs={"task": task}, return_timestamps=True)
text = out["text"]
return [[transcript] for transcript in text.split(".") if transcript], text
with gr.Blocks() as demo:
with gr.Row():
with gr.Column():
audio_input = gr.Audio(source="upload", type="filepath", label="Upload Audio")
task_input = gr.Dropdown(choices=["transcribe", "translate"], value="transcribe", label="Task")
submit_button = gr.Button("Transcribe")
with gr.Column():
output_text = gr.Dataframe(label="Transcripts")
output_full_text = gr.Textbox(label="Full Text")
submit_button.click(
transcribe_simple,
inputs=[audio_input, task_input],
outputs=[output_text, output_full_text],
)
demo.launch()
|