Create app.py

app.py

@@ -0,0 +1,34 @@
+import gradio as gr
+import time
+from transformers import pipeline, AutoModelForSequenceClassification, AutoTokenizer
+
+# ASR pipeline
+asr_pipeline = pipeline("automatic-speech-recognition", model="facebook/wav2vec2-base-960h")
+
+# Load classifier model and tokenizer
+classifier_model = AutoModelForSequenceClassification.from_pretrained("Ngadou/bert-sms-spam-dectector")
+classifier_tokenizer = AutoTokenizer.from_pretrained("Ngadou/bert-sms-spam-dectector")
+
+def classify_audio(audio):
+    # Transcribe the audio to text
+    text = asr_pipeline(audio)["text"]
+
+    # Tokenize the text and feed it to the model
+    inputs = classifier_tokenizer.encode_plus(text, return_tensors="pt")
+    outputs = classifier_model(**inputs)
+
+    # Get the prediction (0 = ham, 1 = spam)
+    prediction = outputs.logits.argmax(dim=1).item()
+
+    # Return the transcription and the prediction as a dictionary
+    return text, "Scam" if prediction == 1 else "Safe Message"
+
+gr.Interface(
+    fn=classify_audio, 
+    inputs=gr.inputs.Audio(source="upload", type="filepath"),
+    outputs=[
+        gr.outputs.Textbox(label="Transcription"),
+        gr.outputs.Textbox(label="Classification"),
+    ],
+    live=True
+).launch(share=True)