appp

app.py

@@ -0,0 +1,41 @@
+import streamlit as st
+import torch
+import librosa
+from transformers import HubertForSequenceClassification, Wav2Vec2FeatureExtractor
+from transformers import pipeline
+
+# Title of the app
+st.title("Emotion Recognition from Speech")
+
+# Upload audio file
+uploaded_file = st.file_uploader("Choose an audio file...", type=["wav"])
+
+# Load the model and feature extractor
+model = HubertForSequenceClassification.from_pretrained("superb/hubert-large-superb-er")
+feature_extractor = Wav2Vec2FeatureExtractor.from_pretrained("superb/hubert-large-superb-er")
+classifier = pipeline("audio-classification", model="superb/hubert-large-superb-er")
+
+if uploaded_file is not None:
+    # Load and preprocess audio file
+    speech, sr = librosa.load(uploaded_file, sr=16000, mono=True)
+    
+    # Display audio player
+    st.audio(uploaded_file, format='audio/wav')
+    
+    # Process the audio
+    inputs = feature_extractor(speech, sampling_rate=16000, padding=True, return_tensors="pt")
+    
+    # Predict emotion
+    with torch.no_grad():
+        logits = model(**inputs).logits
+        predicted_ids = torch.argmax(logits, dim=-1)
+        labels = [model.config.id2label[_id] for _id in predicted_ids.tolist()]
+    
+    # Display the result
+    st.write("Predicted Emotion:", labels[0])
+    
+    # Alternatively using the pipeline
+    results = classifier(uploaded_file, top_k=5)
+    st.write("Top 5 Predicted Emotions:")
+    for result in results:
+        st.write(f"{result['label']}: {result['score']:.4f}")

requirements.txt

@@ -0,0 +1,5 @@
+
+streamlit
+torch
+librosa
+transformers