many bug fix

AsahiMoon · Nov 21, 2022 · b057f1f · b057f1f
1 parent dc3a824
commit b057f1f
Show file tree

Hide file tree

Showing 7 changed files with 24 additions and 26 deletions.
diff --git a/infer_tools/infer_tool.py b/infer_tools/infer_tool.py
@@ -119,7 +119,6 @@ def __init__(self, project_name, config_name, hubert_gpu, model_path):
         self.load_ckpt()
         self.model.cuda()
         hparams['hubert_gpu'] = hubert_gpu
-        hparams['use_uv'] = True
         self.hubert = Hubertencoder(hparams['hubert_path'])
         self.pe = PitchExtractor().cuda()
         utils.load_ckpt(self.pe, hparams['pe_ckpt'], 'model', strict=True)
@@ -151,7 +150,6 @@ def diff_infer():
         batch['mel2ph_pred'] = outputs['mel2ph']
         batch['f0_gt'] = denorm_f0(batch['f0'], batch['uv'], hparams)
         if use_pe:
-            hparams['use_uv'] = True
             batch['f0_pred'] = self.pe(outputs['mel_out'])['f0_denorm_pred'].detach()
         else:
             batch['f0_pred'] = outputs.get('f0_denorm')

diff --git a/inference.ipynb b/inference.ipynb
diff --git a/modules/fastspeech/pe.py b/modules/fastspeech/pe.py
@@ -142,7 +142,7 @@ def forward(self, mel_input=None):
         ret['pitch_pred'] = pitch_pred = self.pitch_predictor(mel_hidden)
 
         pitch_padding = mel_input.abs().sum(-1) == 0
-        use_uv = hparams['pitch_type'] == 'frame' and hparams['use_uv']
+        use_uv = hparams['pitch_type'] == 'frame' #and hparams['use_uv']
         ret['f0_denorm_pred'] = denorm_f0(
             pitch_pred[:, :, 0], (pitch_pred[:, :, 1] > 0) if use_uv else None,
             hparams, pitch_padding=pitch_padding)

diff --git a/preprocessing/hubertinfer.py b/preprocessing/hubertinfer.py
@@ -25,7 +25,7 @@ def __init__(self, pt_path='checkpoints/hubert/hubert_soft.pt'):
             else:
                 self.use_gpu = True
             self.dev = torch.device("cuda" if self.use_gpu and torch.cuda.is_available() else "cpu")
-            self.hbt_model = hubert_soft(str(pt_path))
+            self.hbt_model = hubert_soft(str(pt_path)).to(self.dev)
 
     def encode(self, wav_path):
         if isinstance(wav_path, BytesIO):

diff --git a/test_output.wav b/test_output.wav
diff --git a/training/config.yaml b/training/config.yaml
@@ -333,7 +333,7 @@ use_pos_embed: true
 use_spk_embed: false
 use_spk_id: false
 use_split_spk_id: false
-use_uv: true
+use_uv: false
 use_vec: false
 use_var_enc: false
 val_check_interval: 2000

diff --git a/training/config_nsf.yaml b/training/config_nsf.yaml
@@ -175,7 +175,7 @@ use_pos_embed: true
 use_spk_embed: false
 use_spk_id: false
 use_split_spk_id: false
-use_uv: true
+use_uv: false
 use_vec: false
 use_var_enc: false
 val_check_interval: 2000