Update app.py

app.py

@@ -3,30 +3,42 @@ import gradio as gr  # Add this import statement
 
 subprocess.run(["python", "-m", "pip", "install", "--upgrade", "pip"])
 subprocess.run(["pip", "install", "gradio", "--upgrade"])
-subprocess.run(["pip", "install", "datasets"])
 subprocess.run(["pip", "install", "transformers"])
-subprocess.run(["pip", "install", "librosa", "soundfile"])
 subprocess.run(["pip", "install", "torch", "torchvision", "torchaudio", "-f", "https://download.pytorch.org/whl/torch_stable.html"])
 
+# Install necessary libraries
+!pip install gradio torch torchaudio
+
 import gradio as gr
-from transformers import WhisperProcessor, WhisperForConditionalGeneration
-import numpy as np
+import torchaudio
+from transformers import pipeline
 
-# Load model and processor
-processor = WhisperProcessor.from_pretrained("openai/whisper-small")
-model = WhisperForConditionalGeneration.from_pretrained("openai/whisper-small")
-forced_decoder_ids = processor.get_decoder_prompt_ids(language="italian", task="transcribe")
+# Load the Whispy/Whisper Italian ASR model
+whisper_italian_asr = pipeline("whisper-italian")
 
+# Define the ASR function
 def transcribe_audio(audio):
-    # Assuming sampling_rate is known
-    sampling_rate = 16000  # Change this to the actual sampling rate of your audio
-
-    # Ensure to pass the sampling_rate parameter
-    input_features = processor(audio, sampling_rate=sampling_rate, return_tensors="pt").input_features
-    predicted_ids = model.generate(input_features)
-    transcription = processor.batch_decode(predicted_ids, skip_special_tokens=True)
-    return transcription[0]
-
-# Create Gradio interface
-audio_input = gr.Audio()
-gr.Interface(fn=transcribe_audio, inputs=audio_input, outputs="text").launch()
+    # Save the audio file
+    torchaudio.save("user_audio.wav", audio.squeeze().numpy(), 16000)
+
+    # Load the saved audio file
+    user_audio, _ = torchaudio.load("user_audio.wav", normalize=True)
+
+    # Perform ASR using the Whispy/Whisper Italian model
+    transcription = whisper_italian_asr(user_audio.numpy())
+
+    return transcription[0]["transcription"]
+
+# Create the Gradio interface
+audio_input = gr.Audio(preprocess=torchaudio.transforms.Resample(orig_freq=44100, new_freq=16000))
+
+iface = gr.Interface(
+    fn=transcribe_audio,
+    inputs=audio_input,
+    outputs="text",
+    live=True,
+    interpretation="default"
+)
+
+# Launch the Gradio app
+iface.launch(share=True)