Update app.py

app.py

@@ -2,42 +2,33 @@ import os
 import requests
 import gradio as gr
 import moviepy.editor as mp
+from TTS.api import TTS
 import torch
 import assemblyai as aai
-from xtts_v2 import model.pth
-
-# Import specific model components
-from TTS.tts.configs.xtts_config import XttsConfig
-from TTS.tts.models.xtts import Xtts
-
-# Define paths for model and configuration files
-# model_path = "/xtts_v2"
-#config_path = os.path.join(model_path, "/config.json")
-# checkpoint_path = model_path
-
-# Initialize and load the XTTS model
-config = XttsConfig("/xtts_v2/config.json")
-model = Xtts.init_from_config(config)
-model.load_checkpoint(config, checkpoint_dir="/xtts_v2", eval=True)
-model.cuda()  # Move model to GPU if available
-
-def synthesize_text(text, speaker_wav, language):
-    try:
-        outputs = model.synthesize(
-            text,
-            config,
-            speaker_wav=speaker_wav,
-            gpt_cond_len=3,
-            language=language
-        )
-        return outputs
-    except Exception as e:
-        print(f"Error during synthesis: {e}")
-        raise
+
+# Download necessary models if not already present
+model_files = {
+    "wav2lip.pth": "https://github.com/justinjohn0306/Wav2Lip/releases/download/models/wav2lip.pth",
+    "wav2lip_gan.pth": "https://github.com/justinjohn0306/Wav2Lip/releases/download/models/wav2lip_gan.pth",
+    "resnet50.pth": "https://github.com/justinjohn0306/Wav2Lip/releases/download/models/resnet50.pth",
+    "mobilenet.pth": "https://github.com/justinjohn0306/Wav2Lip/releases/download/models/mobilenet.pth",
+    "s3fd.pth": "https://www.adrianbulat.com/downloads/python-fan/s3fd-619a316812.pth"
+}
+
+for filename, url in model_files.items():
+    file_path = os.path.join("checkpoints" if "pth" in filename else "face_detection", filename)
+    if not os.path.exists(file_path):
+        print(f"Downloading {filename}...")
+        r = requests.get(url)
+        with open(file_path, 'wb') as f:
+            f.write(r.content)
+
+# Initialize TTS model without prompts
+tts = TTS("tts_models/multilingual/multi-dataset/xtts_v2", gpu=True, progress_bar=False)
 
 # Translation class
-class Translation:
-    def _init_(self, video_path, original_language, target_language):
+class translation:
+    def init(self, video_path, original_language, target_language):
         self.video_path = video_path
         self.original_language = original_language
         self.target_language = target_language
@@ -78,18 +69,8 @@ class Translation:
         return translation
 
     def generate_audio(self, translated_text):
-        try:
-            synthesized_audio = synthesize_text(
-                translated_text,
-                speaker_wav='output_audio.wav',
-                language=self.tran_code
-            )
-            with open("output_synth.wav", "wb") as f:
-                f.write(synthesized_audio)
-            return "output_synth.wav"
-        except Exception as e:
-            print(f"Error generating audio: {e}")
-            raise
+        tts.tts_to_file(text=translated_text, speaker_wav='output_audio.wav', file_path="output_synth.wav", language=self.tran_code)
+        return "output_synth.wav"
 
     def translate_video(self):
         audio_path = self.extract_audio()
@@ -106,7 +87,7 @@ class Translation:
 
 # Gradio Interface
 def app(video_path, original_language, target_language):
-    translator = Translation(video_path, original_language, target_language)
+    translator = translation(video_path, original_language, target_language)
     video_file = translator.translate_video()
     return video_file
 
@@ -120,4 +101,4 @@ interface = gr.Interface(
     outputs=gr.Video(label="Translated Video")
 )
 
-interface.launch(share=True)  # Optional: Set share=True to create a public link
+interface.launch()