Deploy Music Generation Studio - 2025-12-12 18:06

backend/services/diffrhythm_service.py

@@ -123,16 +123,22 @@ class DiffRhythmService:
             
             # Load tokenizer
             from g2p.g2p_generation import chn_eng_g2p
-            vocab_path = os.path.join(self.model_path, "vocab.json")
-            if not os.path.exists(vocab_path):
-                # Download vocab
+            
+            # Look for vocab.json in the cloned DiffRhythm2 source
+            diffrhythm2_src = Path(__file__).parent.parent.parent / "models" / "diffrhythm2_source"
+            vocab_path = diffrhythm2_src / "g2p" / "g2p" / "vocab.json"
+            
+            if not vocab_path.exists():
+                # Fallback: try downloading from HF Hub
+                logger.warning(f"vocab.json not found at {vocab_path}, trying to download from HF Hub")
                 vocab_path = hf_hub_download(
                     repo_id=repo_id,
-                    filename="g2p/g2p/vocab.json",
+                    filename="vocab.json",
                     local_dir=self.model_path,
                     local_files_only=False,
                 )
             
+            logger.info(f"Loading vocab from: {vocab_path}")
             with open(vocab_path, 'r') as f:
                 phone2id = json.load(f)['vocab']