Update app.py

app.py

@@ -1,12 +1,20 @@
 
 
 
+import whisper
+
 
 
 import torch
 from transformers import AutoTokenizer, AutoModelForCausalLM
 import gradio as gr
 
+
+Asr_model = whisper.load_model("base")
+Asr_model.device
+
+
+
 device = "cuda" if torch.cuda.is_available() else "cpu"
 
 model_name = "ai4bharat/Airavata"
@@ -52,10 +60,40 @@ def inference(input_prompt, model, tokenizer):
     return output_text
 
 
-def chat_interface(message,history):
+
+def transcribe(audio):
+    
+    #time.sleep(3)
+    # load audio and pad/trim it to fit 30 seconds
+    audio = whisper.load_audio(audio)
+    audio = whisper.pad_or_trim(audio)
+
+    # make log-Mel spectrogram and move to the same device as the model
+    mel = whisper.log_mel_spectrogram(audio).to(model.device)
+
+    # detect the spoken language
+    _, probs = model.detect_language(mel)
+    print(f"Detected language: {max(probs, key=probs.get)}")
+
+    # decode the audio
+    options = whisper.DecodingOptions()
+    result = whisper.decode(model, mel, options)
+    return result.text
+
+
+def chat_interface(audio):
+    message = transcribe(audio)
     outputs = inference(message, model, tokenizer)
     return outputs
 
 
-chat_interface = gr.ChatInterface(chat_interface, title="CAMAI")
-chat_interface.launch()
+gr.Interface(
+    title = 'CAMAI - Centralized Actionable Multimodal Agri Assistant on Edge Intelligence for Farmers ', 
+    fn=chat_interface, 
+    inputs=[
+        gr.inputs.Audio(source="microphone", type="filepath")
+    ],
+    outputs=[
+        "textbox"
+    ],
+    live=True).launch()