Update app.py

app.py

@@ -1,6 +1,5 @@
 import gradio as gr
 from TTS.utils.synthesizer import Synthesizer
-from TTS.tts.configs.vits_config import VitsConfig
 from huggingface_hub import hf_hub_download
 import os
 import json
@@ -10,22 +9,27 @@ def load_eng_model():
     repo_id = "E-motionAssistant/text-to-speech-VITS-english"
     print(f"--- Downloading English Model ---")
     
+    # 1. Download original files
     model_path = hf_hub_download(repo_id=repo_id, filename="best_model.pth")
     config_path = hf_hub_download(repo_id=repo_id, filename="config.json")
 
-    # 1. Load and Fix the Config Object
-    config = VitsConfig()
-    config.load_json(config_path)
+    # 2. Read the config file and manually change the number to 137
+    with open(config_path, 'r') as f:
+        config_dict = json.load(f)
     
-    print("Force-setting num_chars to 137...")
-    config.model_args.num_chars = 137
+    print("Fixing character count: 131 -> 137")
+    config_dict["model_args"]["num_chars"] = 137
     
-    # 2. Use Synthesizer directly instead of the TTS wrapper
-    # This bypasses the logic that causes the 131 vs 137 mismatch
-    print("Building Synthesizer...")
+    # 3. Save this as a NEW file (a physical path the library will accept)
+    fixed_config_path = os.path.join(os.getcwd(), "fixed_config.json")
+    with open(fixed_config_path, 'w') as f:
+        json.dump(config_dict, f)
+
+    # 4. Initialize the Synthesizer using the PATH to our new file
+    print("Building Synthesizer with fixed_config.json path...")
     syn = Synthesizer(
         tts_checkpoint=model_path,
-        tts_config_path=config, # We pass the FIXED object directly
+        tts_config_path=fixed_config_path, # This is now a STRING path
         use_cuda=False
     )
     
@@ -35,7 +39,6 @@ def load_eng_model():
 # --- Global Initialization ---
 print("Starting English TTS Startup...")
 try:
-    # eng_tts is now a Synthesizer object
     eng_tts = load_eng_model()
     print("--- SUCCESS: ENGLISH MODEL LOADED ---")
 except Exception as e:
@@ -43,18 +46,14 @@ except Exception as e:
     eng_tts = None
 
 def generate_voice(text):
-    if eng_tts is None:
+    if not eng_tts:
         return None
     try:
-        output_path = os.path.join(os.getcwd(), "en_output.wav")
-        
-        # Synthesizer uses 'tts' instead of 'tts_to_file'
-        # It returns a list of audio samples
+        output_path = os.path.join(os.getcwd(), "output.wav")
+        # Generate the audio samples
         wav = eng_tts.tts(text=str(text))
-        
-        # We save the wav samples to a file
+        # Save samples to a .wav file
         eng_tts.save_wav(wav, output_path)
-        
         return output_path
     except Exception as e:
         print(f"Generation Error: {e}")
@@ -63,9 +62,9 @@ def generate_voice(text):
 # Gradio Interface
 demo = gr.Interface(
     fn=generate_voice,
-    inputs=gr.Textbox(label="Input English Text"),
+    inputs=gr.Textbox(label="Input English Text", placeholder="Hello world!"),
     outputs=gr.Audio(label="Synthesized Speech", type="filepath"),
-    title="English VITS TTS (Direct Synth)"
+    title="English VITS TTS"
 )
 
 if __name__ == "__main__":