replacing transformers pipeline with whisper python library

app.py

@@ -1,14 +1,36 @@
+import whisper
 import gradio as gr
-from transformers import pipeline
 
-p = pipeline("automatic-speech-recognition", model='openai/whisper-small')
+model = whisper.load_model("small")
 
 def transcribe(audio):
-    text = p(audio)["text"]
-    return text
+    
+    #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(fp16 = False)
+    result = whisper.decode(model, mel, options)
+    return result.text
+    
+    
+ 
 gr.Interface(
+    title = 'OpenAI Whisper ASR Gradio Web UI', 
     fn=transcribe, 
-    inputs=gr.Audio(source="microphone", type="filepath"), 
-    outputs="text").launch()
-
+    inputs=[
+        gr.inputs.Audio(source="microphone", type="filepath")
+    ],
+    outputs=[
+        "textbox"
+    ],
+    live=True).launch()