Update app.py

app.py

@@ -1,45 +1,39 @@
 from fastapi import FastAPI, File, UploadFile
 from fastapi.middleware.cors import CORSMiddleware
-from pydantic import BaseModel
 from transformers import pipeline
 import uvicorn
 import tempfile
+import torchaudio
 
-# Initialize FastAPI
 app = FastAPI()
 
-# Enable CORS for all origins (so Render or any client can access it)
+# Allow CORS
 app.add_middleware(
     CORSMiddleware,
     allow_origins=["*"],
-    allow_credentials=True,
     allow_methods=["*"],
     allow_headers=["*"],
 )
 
-# Load the pretrained speech emotion recognition pipeline
-emotion_pipeline = pipeline(
-    "audio-classification",
-    model="ehcalabres/wav2vec2-lg-xlsr-en-speech-emotion-recognition"
-)
-
-# Health check route
-@app.get("/")
-def read_root():
-    return {"message": "HF Space is live!"}
+# Load model
+pipe = pipeline("audio-classification", model="superb/wav2vec2-base-superb-er")
 
-# Predict route
 @app.post("/predict")
-async def predict_emotion(file: UploadFile = File(...)):
+async def predict(file: UploadFile = File(...)):
     try:
-        # Save the uploaded audio file to a temporary location
-        with tempfile.NamedTemporaryFile(delete=False) as tmp:
+        # Save uploaded file to a temp file
+        with tempfile.NamedTemporaryFile(delete=False, suffix=".wav") as tmp:
             tmp.write(await file.read())
             tmp_path = tmp.name
 
-        # Run emotion prediction
-        result = emotion_pipeline(tmp_path)
-        top_emotion = result[0]['label']
+        # Load and preprocess audio
+        waveform, sample_rate = torchaudio.load(tmp_path)
+
+        # Get prediction
+        result = pipe(tmp_path)
+
+        # Get top prediction label
+        top_emotion = result[0]["label"].lower()
 
         return {"emotion": top_emotion}