v0.2.1 Fixed #545

pyproject.toml

@@ -4,7 +4,7 @@ build-backend = "setuptools.build_meta"
 
 [project]
 name = "f5-tts"
-version = "0.2.0"
+version = "0.2.1"
 description = "F5-TTS: A Fairytaler that Fakes Fluent and Faithful Speech with Flow Matching"
 readme = "README.md"
 license = {text = "MIT License"}

src/f5_tts/configs/E2TTS_Base_train.yaml

@@ -33,8 +33,9 @@ model:
     win_length: 1024
     n_fft: 1024
     mel_spec_type: vocos  # 'vocos' or 'bigvgan'
-    is_local_vocoder: False  # use local offline vocoder ckpt or not
-    local_vocoder_path: None  # path to local vocoder
+  vocoder:
+    is_local: False  # use local offline ckpt or not
+    local_path: None  # local vocoder path
 
 ckpts:
   logger: wandb  # wandb | tensorboard | None

src/f5_tts/configs/E2TTS_Small_train.yaml

@@ -33,8 +33,9 @@ model:
     win_length: 1024
     n_fft: 1024
     mel_spec_type: vocos  # 'vocos' or 'bigvgan'
-    is_local_vocoder: False
-    local_vocoder_path: None
+  vocoder:
+    is_local: False  # use local offline ckpt or not
+    local_path: None  # local vocoder path
 
 ckpts:
   logger: wandb  # wandb | tensorboard | None

src/f5_tts/configs/F5TTS_Base_train.yaml

@@ -35,8 +35,9 @@ model:
     win_length: 1024
     n_fft: 1024
     mel_spec_type: vocos  # 'vocos' or 'bigvgan'
-    is_local_vocoder: False  # use local offline vocoder ckpt or not
-    local_vocoder_path: None  # local vocoder path
+  vocoder:
+    is_local: False  # use local offline ckpt or not
+    local_path: None  # local vocoder path
 
 ckpts:
   logger: wandb  # wandb | tensorboard | None

src/f5_tts/configs/F5TTS_Small_train.yaml

@@ -35,8 +35,9 @@ model:
     win_length: 1024
     n_fft: 1024
     mel_spec_type: vocos  # 'vocos' or 'bigvgan'
-    is_local_vocoder: False
-    local_vocoder_path: None
+  vocoder:
+    is_local: False  # use local offline ckpt or not
+    local_path: None  # local vocoder path
 
 ckpts:
   logger: wandb  # wandb | tensorboard | None

src/f5_tts/train/train.py

@@ -59,8 +59,8 @@ def main(cfg):
         log_samples=True,
         bnb_optimizer=cfg.optim.bnb_optimizer,
         mel_spec_type=mel_spec_type,
-        is_local_vocoder=cfg.model.mel_spec.is_local_vocoder,
-        local_vocoder_path=cfg.model.mel_spec.local_vocoder_path,
+        is_local_vocoder=cfg.model.vocoder.is_local,
+        local_vocoder_path=cfg.model.vocoder.local_path,
     )
 
     train_dataset = load_dataset(cfg.datasets.name, tokenizer, mel_spec_kwargs=cfg.model.mel_spec)