Create app.py

app.py

@@ -0,0 +1,38 @@
+# Load model directly
+from transformers import AutoProcessor, AutoModelForSpeechSeq2Seq
+import torchaudio
+import streamlit as st
+
+processor = AutoProcessor.from_pretrained("mohammed/whisper-small-arabic-cv-11")
+model = AutoModelForSpeechSeq2Seq.from_pretrained("mohammed/whisper-small-arabic-cv-11")
+
+st.title("Arabic Whisper model v2")
+
+audio_file = st.file_uploader("Upload audio", type=["mp3", "wav", "m4a"])
+
+if st.sidebar.button("Trascribe Audio"):
+    if audio_file is not None:
+        st.sidebar.success("Transcribing audio") # on success audio file
+        
+        audio_tensor, sample_rate = torchaudio.load(audio_file)
+        
+        if sample_rate != 16000:
+            resampler = torchaudio.transforms.Resample(orig_freq=sample_rate, new_freq=16000)
+            audio_tensor = resampler(audio_tensor)
+            
+        audio_np = audio_tensor.squeeze().numpy()
+        
+        # processing audio
+        inputs = processor(audio_np, sample_rate=16000, return_tensors="pt")
+        
+        # generating transcript
+        generated_ids = model.generate(inputs["input_features"])
+        
+        transcription = processor.batch_decode(generated_ids, skip_special_tokens=True)
+        
+        # display transcription
+        st.sidebar.success("Transcription is complete")
+        st.text(transcription[0])
+        
+    else:
+        st.sidebar.error("Please upload a valid audio file")