Create app.py

app.py

@@ -0,0 +1,52 @@
+from flask import Flask, render_template, request, jsonify
+import torch
+from transformers import pipeline
+import gradio as gr
+
+app = Flask(__name__)
+
+# Load the automatic speech recognition model
+pipe = pipeline("automatic-speech-recognition",
+               "openai/whisper-large-v3",
+               torch_dtype=torch.float16,
+               device="cuda:0")
+
+# Load the emotion classification model
+emotion_classifier = pipeline(
+    "text-classification",
+    model="j-hartmann/emotion-english-distilroberta-base",
+    return_all_scores=True
+)
+
+def transcribe(audio_file, task):
+    if audio_file is None:
+        return "Please upload or record an audio file."
+
+    # Check if the audio file is in bytes format (drag-and-drop file)
+    if isinstance(audio_file, bytes):
+        text = pipe(audio_file, generate_kwargs={"task": task}, return_timestamps=True)["text"]
+    else:
+        # Handle the case where the file is uploaded using the file uploader
+        text = pipe(audio_file.name, generate_kwargs={"task": task}, return_timestamps=True)["text"]
+
+    return text
+
+@app.route('/')
+def index():
+    return render_template('index.html')
+
+@app.route('/transcribe', methods=['POST'])
+def transcribe_endpoint():
+    audio_file = request.files.get('audio_file')
+    task = request.form.get('task')
+    text = transcribe(audio_file, task)
+    return jsonify({'text': text})
+
+@app.route('/classify_emotion', methods=['POST'])
+def classify_emotion_endpoint():
+    text = request.form.get('text')
+    result = emotion_classifier(text)
+    return jsonify(result)
+
+if __name__ == '__main__':
+    app.run(debug=True)