Update app.py

app.py

@@ -1,45 +1,15 @@
-import os
 import gradio as gr
-import soundfile as sf
-from transformers import AutoTokenizer, AutoModelForSeq2SeqLM
-from gtts import gTTS
+import whisper
 
-# Load the translation model
-translation_tokenizer = AutoTokenizer.from_pretrained("alirezamsh/small100")
-translation_model = AutoModelForSeq2SeqLM.from_pretrained("alirezamsh/small100")
+model = whisper.load_model("base")
 
-# Available target languages
-available_languages = {
-    'Russian': 'ru',
-    'Spanish': 'es',
-    'English': 'en',
-    'Greek': 'gr'
-}
+def transcribe(audio):
+    audio = whisper.pad_or_trim(audio)
+    mel = whisper.log_mel_spectrogram(audio).to(model.device)
+    options = whisper.DecodingOptions()
+    result = whisper.decode(model, mel, options)
+    return result.text
 
-# Function to translate the audio
-def translate_audio(audio_file, target_language):
-    to_lang = available_languages[target_language]
-
-    # Load audio
-    audio, sample_rate = sf.read(audio_file.name)
-
-    # Translate the text
-    translation_tokenizer.src_lang = to_lang
-    encoded_bg = translation_tokenizer(audio, return_tensors="pt", padding=True, truncation=True)
-    generated_tokens = translation_model.generate(**encoded_bg)
-    translated_audio = translation_tokenizer.batch_decode(generated_tokens, skip_special_tokens=True)
-
-    # Save translated audio
-    output_file = "translated_audio.wav"
-    sf.write(output_file, translated_audio, sample_rate)
-
-    return output_file
-
-# Gradio interface
-audio_input = gr.inputs.Audio(label="Upload audio file")
-language_dropdown = gr.inputs.Dropdown(choices=list(available_languages.keys()), label="Select Target Language")
-audio_output = gr.outputs.Audio(label="Translated audio file")
-
-iface = gr.Interface(fn=translate_audio, inputs=[audio_input, language_dropdown], outputs=audio_output, title="Audio Translation Demo")
+iface = gr.Interface(fn=transcribe, inputs="audio", outputs="text")
 iface.launch()