fix lint

cosyvoice/flow/decoder.py

@@ -74,7 +74,7 @@ class ConditionalDecoder(nn.Module):
             )
             self.down_blocks.append(nn.ModuleList([resnet, transformer_blocks, downsample]))
 
-        for i in range(num_mid_blocks):
+        for _ in range(num_mid_blocks):
             input_channel = channels[-1]
             out_channels = channels[-1]
             resnet = ResnetBlock1D(dim=input_channel, dim_out=output_channel, time_emb_dim=time_embed_dim)
@@ -126,7 +126,6 @@ class ConditionalDecoder(nn.Module):
         self.final_proj = nn.Conv1d(channels[-1], self.out_channels, 1)
         self.initialize_weights()
 
-
     def initialize_weights(self):
         for m in self.modules():
             if isinstance(m, nn.Conv1d):

cosyvoice/flow/flow.py

@@ -33,8 +33,13 @@ class MaskedDiffWithXvec(torch.nn.Module):
                  encoder: torch.nn.Module = None,
                  length_regulator: torch.nn.Module = None,
                  decoder: torch.nn.Module = None,
-                 decoder_conf: Dict = {'in_channels': 240, 'out_channel': 80, 'spk_emb_dim': 80, 'n_spks': 1, 'cfm_params': DictConfig({'sigma_min': 1e-06, 'solver': 'euler', 't_scheduler': 'cosine', 'training_cfg_rate': 0.2, 'inference_cfg_rate': 0.7, 'reg_loss_type': 'l1'}), 'decoder_params': {'channels': [256, 256], 'dropout': 0.0, 'attention_head_dim': 64, 'n_blocks': 4, 'num_mid_blocks': 12, 'num_heads': 8, 'act_fn': 'gelu'}},
-                 mel_feat_conf: Dict = {'n_fft': 1024, 'num_mels': 80, 'sampling_rate': 22050, 'hop_size': 256, 'win_size': 1024, 'fmin': 0, 'fmax': 8000}):
+                 decoder_conf: Dict = {'in_channels': 240, 'out_channel': 80, 'spk_emb_dim': 80, 'n_spks': 1,
+                                       'cfm_params': DictConfig({'sigma_min': 1e-06, 'solver': 'euler', 't_scheduler': 'cosine',
+                                                                 'training_cfg_rate': 0.2, 'inference_cfg_rate': 0.7, 'reg_loss_type': 'l1'}),
+                                       'decoder_params': {'channels': [256, 256], 'dropout': 0.0, 'attention_head_dim': 64,
+                                                          'n_blocks': 4, 'num_mid_blocks': 12, 'num_heads': 8, 'act_fn': 'gelu'}},
+                 mel_feat_conf: Dict = {'n_fft': 1024, 'num_mels': 80, 'sampling_rate': 22050,
+                                        'hop_size': 256, 'win_size': 1024, 'fmin': 0, 'fmax': 8000}):
         super().__init__()
         self.input_size = input_size
         self.output_size = output_size

cosyvoice/flow/flow_matching.py

@@ -15,6 +15,7 @@ import torch
 import torch.nn.functional as F
 from matcha.models.components.flow_matching import BASECFM
 
+
 class ConditionalCFM(BASECFM):
     def __init__(self, in_channels, cfm_params, n_spks=1, spk_emb_dim=64, estimator: torch.nn.Module = None):
         super().__init__(