Create app.py

app.py

@@ -0,0 +1,26 @@
+# pip install transformers datasets torchaudio soundfile
+from transformers import pipeline
+import torchaudio
+
+# 1. Whisper ASR model (Vietnamese)
+asr = pipeline("automatic-speech-recognition", model="openai/whisper-small", device=-1)
+
+# 2. Emotion classification model (Vietnamese)
+emo_clf = pipeline("text-classification", model="bkai-foundation-models/vietnamese-emotion", top_k=None)
+
+# 3. Pipeline: audio -> transcript -> emotion
+def predict_emotion(audio_path):
+    # Chuyển audio thành text
+    transcript = asr(audio_path)["text"]
+    # Phân loại cảm xúc
+    emotions = emo_clf(transcript)[0]
+    # Sắp xếp theo độ tin cậy
+    emotions = sorted(emotions, key=lambda x: x['score'], reverse=True)
+    return transcript, emotions
+
+# Demo
+if __name__ == "__main__":
+    audio_file = "sample_vi.wav"  # file giọng nói tiếng Việt
+    text, emo = predict_emotion(audio_file)
+    print("Transcript:", text)
+    print("Emotion prediction:", emo)