Create app.py

app.py

@@ -0,0 +1,74 @@
+import gradio as gr
+import torch
+import whisper
+import warnings
+warnings.filterwarnings('ignore')
+from transformers import pipeline
+import os
+
+MODEL_NAME = "openai/whisper-small"
+BATCH_SIZE = 8
+
+device = 0 if torch.cuda.is_available() else "cpu"
+
+pipe = pipeline(
+    task="automatic-speech-recognition",
+    model=MODEL_NAME,
+    chunk_length_s=30,
+    device=device)
+
+emotion_classifier = pipeline("text-classification",model='MilaNLProc/xlm-emo-t', return_all_scores=True)
+
+def transcribe(microphone, file_upload, task):
+    output = ""
+    if (microphone is not None) and (file_upload is not None):
+        warn_output = (
+            "WARNING: You've uploaded an audio file and used the microphone. "
+            "The recorded file from the microphone will be used and the uploaded audio will be discarded.\n"
+        )
+
+    elif (microphone is None) and (file_upload is None):
+        raise gr.Error("You have to either use the microphone or upload an audio file")
+
+    file = microphone if microphone is not None else file_upload
+
+    text = pipe(file, batch_size=BATCH_SIZE, generate_kwargs={"task": task})["text"]
+
+    return output + text
+
+def translate_and_classify(audio):
+    text_result = transcribe(audio, None, "transcribe")    
+    emotion = emotion_classifier(text_result)
+    detected_emotion = {}
+    for emotion in emotion[0]:
+        detected_emotion[emotion["label"]] = emotion["score"]
+    return text_result, detected_emotion
+
+with gr.Blocks() as demo:
+
+    gr.Markdown(
+    """    # Emotion Detection from Speech
+    
+            ##### Detection of anger, sadness, joy, fear in speech using OpenAI Whisper and XLM-RoBERTa 
+            
+            """)
+    
+    with gr.Column():
+        with gr.Tab("Record Audio"):
+            # The 'source' argument is no longer supported, use 'sources' instead
+            audio_input_r = gr.Audio(label = 'Record Audio Input',sources=["microphone"],type="filepath") 
+            transcribe_audio_r = gr.Button('Transcribe')
+        
+        with gr.Tab("Upload Audio as File"):
+            # The 'source' argument is no longer supported, use 'sources' instead
+            audio_input_u = gr.Audio(label = 'Upload Audio',sources=["upload"],type="filepath") 
+            transcribe_audio_u = gr.Button('Transcribe')
+
+        with gr.Row():
+            transcript_output = gr.Textbox(label="Transcription in the language of speech/audio", lines = 3)
+            emotion_output = gr.Label(label = "Detected Emotion")
+    
+    transcribe_audio_r.click(translate_and_classify, inputs = audio_input_r, outputs = [transcript_output,emotion_output])
+    transcribe_audio_u.click(translate_and_classify, inputs = audio_input_u, outputs = [transcript_output,emotion_output])       
+    
+demo.launch(share=True)