Fix: Download models from Harshil748/VoiceAPI-Models

download_models.py

@@ -1,104 +1,44 @@
 #!/usr/bin/env python3
 """
 Download models from HuggingFace at build time.
-Uses SYSPIN's hosted models to avoid storage limits.
+Downloads from Harshil748/VoiceAPI-Models repo.
 """
 import os
 from pathlib import Path
-from huggingface_hub import snapshot_download, hf_hub_download
+from huggingface_hub import snapshot_download
 
 MODELS_DIR = Path("models")
-
-# SYSPIN HuggingFace repos for each language
-SYSPIN_REPOS = {
-    # Bengali
-    "bn_female": ("SYSPIN/Bengali_Female_TTS", ["bn_female_vits_30hrs.pt", "chars.txt", "jit_infer.py"]),
-    "bn_male": ("SYSPIN/Bengali_Male_TTS", ["bn_male_vits_30hrs.pt", "chars.txt", "jit_infer.py", "extra.py"]),
-    # Hindi
-    "hi_female": ("SYSPIN/Hindi_Female_TTS", ["hi_female_vits_30hrs.pt", "chars.txt", "jit_infer.py", "extra.py"]),
-    "hi_male": ("SYSPIN/Hindi_Male_TTS", ["hi_male_vits_30hrs.pt", "chars.txt", "jit_infer.py", "extra.py"]),
-    # Marathi
-    "mr_female": ("SYSPIN/Marathi_Female_TTS", ["mr_female_vits_30hrs.pt", "chars.txt", "jit_infer.py", "extra.py"]),
-    "mr_male": ("SYSPIN/Marathi_Male_TTS", ["mr_male_vits_30hrs.pt", "chars.txt", "jit_infer.py", "extra.py"]),
-    # Telugu
-    "te_female": ("SYSPIN/Telugu_Female_TTS", ["te_female_vits_30hrs.pt", "chars.txt", "jit_infer.py", "extra.py"]),
-    "te_male": ("SYSPIN/Telugu_Male_TTS", ["te_male_vits_30hrs.pt", "chars.txt", "jit_infer.py", "extra.py"]),
-    # Kannada
-    "kn_female": ("SYSPIN/Kannada_Female_TTS", ["kn_female_vits_30hrs.pt", "chars.txt", "jit_infer.py", "extra.py"]),
-    "kn_male": ("SYSPIN/Kannada_Male_TTS", ["kn_male_vits_30hrs.pt", "chars.txt", "jit_infer.py", "extra.py"]),
-    # English
-    "en_female": ("SYSPIN/English_Female_TTS", ["en_female_vits_30hrs.pt", "chars.txt", "jit_infer.py", "extra.py"]),
-    "en_male": ("SYSPIN/English_Male_TTS", ["en_male_vits_30hrs.pt", "chars.txt", "jit_infer.py", "extra.py"]),
-    # Chhattisgarhi
-    "hne_female": ("SYSPIN/Chhattisgarhi_Female_TTS", ["ch_female_vits_30hrs.pt", "chars.txt", "jit_infer.py", "extra.py"]),
-    "hne_male": ("SYSPIN/Chhattisgarhi_Male_TTS", ["ch_male_vits_30hrs.pt", "chars.txt", "jit_infer.py", "extra.py"]),
-    # Maithili
-    "mai_female": ("SYSPIN/Maithili_Female_TTS", ["mt_female_vits_30hrs.pt", "chars.txt", "jit_infer.py", "extra.py"]),
-    "mai_male": ("SYSPIN/Maithili_Male_TTS", ["mt_male_vits_30hrs.pt", "chars.txt", "jit_infer.py", "extra.py"]),
-    # Magahi
-    "mag_female": ("SYSPIN/Magahi_Female_TTS", ["mg_female_vits_30hrs.pt", "chars.txt", "jit_infer.py", "extra.py"]),
-    "mag_male": ("SYSPIN/Magahi_Male_TTS", ["mg_male_vits_30hrs.pt", "chars.txt", "jit_infer.py", "extra.py"]),
-    # Bhojpuri (Coqui TTS format)
-    "bho_female": ("SYSPIN/Bhojpuri_Female_TTS", ["checkpoint_340000.pth", "config.json"]),
-    "bho_male": ("SYSPIN/Bhojpuri_Male_TTS", ["checkpoint_200000.pth", "config.json"]),
-}
-
-# Priority voices for demo (download first)
-PRIORITY_VOICES = ["hi_female", "bn_female", "en_female", "mr_female", "te_female", "kn_female"]
-
-
-def download_voice(voice_id: str) -> bool:
-    """Download a single voice model."""
-    if voice_id not in SYSPIN_REPOS:
-        print(f"❌ Unknown voice: {voice_id}")
-        return False
-    
-    repo_id, files = SYSPIN_REPOS[voice_id]
-    voice_dir = MODELS_DIR / voice_id
-    voice_dir.mkdir(parents=True, exist_ok=True)
-    
-    print(f"📥 Downloading {voice_id} from {repo_id}...")
-    
-    try:
-        for filename in files:
-            local_path = voice_dir / filename
-            if local_path.exists():
-                print(f"  ✓ {filename} already exists")
-                continue
-            
-            print(f"  ⬇️ {filename}...")
-            hf_hub_download(
-                repo_id=repo_id,
-                filename=filename,
-                local_dir=voice_dir,
-                local_dir_use_symlinks=False
-            )
-        print(f"✅ {voice_id} downloaded successfully")
-        return True
-    except Exception as e:
-        print(f"❌ Failed to download {voice_id}: {e}")
-        return False
+MODEL_REPO = "Harshil748/VoiceAPI-Models"
 
 
 def download_all_models():
-    """Download all models - priority voices first."""
+    """Download all models from HuggingFace."""
     print("🚀 Starting model downloads...")
-    print(f"   Models directory: {MODELS_DIR.absolute()}")
+    print(f"   Source: {MODEL_REPO}")
+    print(f"   Target: {MODELS_DIR.absolute()}")
     
     MODELS_DIR.mkdir(exist_ok=True)
     
-    # Download priority voices first
-    print("\n📌 Downloading priority voices...")
-    for voice_id in PRIORITY_VOICES:
-        download_voice(voice_id)
-    
-    # Download remaining voices
-    print("\n📦 Downloading remaining voices...")
-    remaining = [v for v in SYSPIN_REPOS.keys() if v not in PRIORITY_VOICES]
-    for voice_id in remaining:
-        download_voice(voice_id)
-    
-    print("\n✨ Model download complete!")
+    try:
+        print("\n�� Downloading all models...")
+        snapshot_download(
+            repo_id=MODEL_REPO,
+            local_dir=MODELS_DIR,
+            local_dir_use_symlinks=False,
+            ignore_patterns=["*.md", ".gitattributes"],
+        )
+        print("✅ All models downloaded successfully!")
+        
+        # List downloaded voices
+        print("\n📦 Downloaded voices:")
+        for voice in sorted(MODELS_DIR.iterdir()):
+            if voice.is_dir():
+                files = list(voice.glob("*"))
+                print(f"  ✓ {voice.name}: {len(files)} files")
+                
+    except Exception as e:
+        print(f"❌ Failed to download models: {e}")
+        raise
 
 
 if __name__ == "__main__":