Create app.py

app.py

@@ -0,0 +1,47 @@
+import gradio as gr
+from transformers import Wav2Vec2ForCTC, Wav2Vec2Tokenizer
+import torch
+import torchaudio
+
+# Load pre-trained model and tokenizer
+model_name = "facebook/wav2vec2-base-960h"
+tokenizer = Wav2Vec2Tokenizer.from_pretrained(model_name)
+model = Wav2Vec2ForCTC.from_pretrained(model_name)
+
+def speech_to_text(audio):
+    # Load audio file
+    waveform, rate = torchaudio.load(audio.name)
+    
+    # Ensure the audio is mono
+    if waveform.shape[0] > 1:
+        waveform = torch.mean(waveform, dim=0, keepdim=True)
+    
+    # Resample to 16000 Hz
+    resampler = torchaudio.transforms.Resample(orig_freq=rate, new_freq=16000)
+    waveform = resampler(waveform)
+    
+    # Tokenize the waveform
+    inputs = tokenizer(waveform.squeeze().numpy(), return_tensors="pt", sampling_rate=16000)
+    
+    # Perform inference
+    with torch.no_grad():
+        logits = model(**inputs).logits
+    
+    # Decode the output
+    predicted_ids = torch.argmax(logits, dim=-1)
+    transcription = tokenizer.batch_decode(predicted_ids)[0]
+    
+    return transcription
+
+# Create Gradio interface
+iface = gr.Interface(
+    fn=speech_to_text,
+    inputs=gr.inputs.Audio(source="microphone", type="file"),
+    outputs="text",
+    live=True,
+    title="Speech to Text",
+    description="Speak into your microphone and get the transcribed text."
+)
+
+# Launch the interface
+iface.launch()