Update app.py

app.py

@@ -4,20 +4,25 @@ import whisper
 from transformers import AutoTokenizer, AutoModelForSeq2SeqLM
 from gtts import gTTS
 
-def translate_voice(file, target_lang):
+def translate_voice(file_obj, target_lang):
+    # Save the temporary file to disk
+    temp_file_path = "temp_audio_file.wav"
+    with open(temp_file_path, "wb") as out_file:
+        out_file.write(file_obj.read())
+
     # Load the model and switch to float32
     model = whisper.load_model("base").float()
 
-# Load the audio
-    audio = whisper.load_audio(file)
+    # Load the audio
+    audio = whisper.load_audio(temp_file_path)
 
-# Pad or trim the audio
+    # Pad or trim the audio
     audio = whisper.pad_or_trim(audio)
 
-# Convert the audio to a log Mel spectrogram and move it to the same device as the model (CPU in your case)
+    # Convert the audio to a log Mel spectrogram and move it to the same device as the model (CPU in your case)
     mel = whisper.log_mel_spectrogram(audio).to(model.device).float()  # convert to full-precision float32
 
-# Proceed with your language detection and decoding
+    # Proceed with your language detection and decoding
     _, probs = model.detect_language(mel)
     options = whisper.DecodingOptions()
     result = whisper.decode(model, mel, options)
@@ -44,14 +49,14 @@ def translate_voice(file, target_lang):
 iface = gr.Interface(
     fn=translate_voice, 
     inputs=[
-        gr.components.File(type="file", label="Your Audio"), 
-        gr.components.Dropdown(choices=['en', 'ru', 'de', 'fr'], label="Target Language")
+        gr.inputs.File(type="file", label="Your Audio"), 
+        gr.inputs.Dropdown(choices=['en', 'ru', 'de', 'fr'], label="Target Language")
     ], 
     outputs=[
-        gr.components.Audio(type="filepath", label="Translated Audio"),
-        gr.components.Textbox(label="Original Text"),
-        gr.components.Textbox(label="Translated Text"),
-        gr.components.Textbox(label="Target Language"),
+        gr.outputs.Audio(type="file", label="Translated Audio"),
+        gr.outputs.Textbox(label="Original Text"),
+        gr.outputs.Textbox(label="Translated Text"),
+        gr.outputs.Textbox(label="Target Language"),
     ]
 )
 iface.launch()