单句推理：RTF性能至少提升 10%

.gitignore

@@ -3,6 +3,7 @@ __pycache__
 *.egg-info
 *.DS_Store
 .idea/
+.vscode/
 checkpoints/*.pth
 checkpoints/*.vocab
 checkpoints/*.model

checkpoints/config.yaml

@@ -19,6 +19,7 @@ gpt:
     use_mel_codes_as_input: true
     mel_length_compression: 1024
     layers: 20
+    activation_function: "gelu_pytorch_tanh"
     number_text_tokens: 12000
     number_mel_codes: 8194
     start_mel_token: 8192

indextts/gpt/model.py

@@ -250,7 +250,7 @@ class LearnedPositionEmbeddings(nn.Module):
         return self.emb(torch.tensor([ind], device=dev)).unsqueeze(0)
 
 
-def build_hf_gpt_transformer(layers, model_dim, heads, max_mel_seq_len, max_text_seq_len, checkpointing):
+def build_hf_gpt_transformer(layers, model_dim, heads, max_mel_seq_len, max_text_seq_len, checkpointing, activation_function):
     """
     GPT-2 implemented by the HuggingFace library.
     """
@@ -261,6 +261,7 @@ def build_hf_gpt_transformer(layers, model_dim, heads, max_mel_seq_len, max_text
                             n_embd=model_dim,
                             n_layer=layers,
                             n_head=heads,
+                            activation_function=activation_function or "gelu_new",
                             gradient_checkpointing=checkpointing,
                             use_cache=not checkpointing)
     gpt = GPT2Model(gpt_config)
@@ -301,7 +302,7 @@ class UnifiedVoice(nn.Module):
                  mel_length_compression=1024, number_text_tokens=256,
                  start_text_token=0, stop_text_token=1, number_mel_codes=8194, start_mel_token=8192, stop_mel_token=8193,
                  train_solo_embeddings=False, use_mel_codes_as_input=True,
-                 checkpointing=True, types=1,
+                 checkpointing=True, types=1, activation_function=None,
                  condition_num_latent=32, condition_type="perceiver", condition_module=None):
         """
         Args:
@@ -365,7 +366,7 @@ class UnifiedVoice(nn.Module):
             self.mel_embedding = MelEncoder(model_dim, resblocks_per_reduction=1)
         self.gpt, self.mel_pos_embedding, self.text_pos_embedding, self.mel_layer_pos_embedding, self.text_layer_pos_embedding = \
             build_hf_gpt_transformer(layers, model_dim, heads, self.max_mel_tokens + 2 + self.max_conditioning_inputs,
-                                     self.max_text_tokens + 2, checkpointing)
+                                     self.max_text_tokens + 2, checkpointing, activation_function)
         if train_solo_embeddings:
             self.mel_solo_embedding = nn.Parameter(torch.randn(1, 1, model_dim) * .02, requires_grad=True)
             self.text_solo_embedding = nn.Parameter(torch.randn(1, 1, model_dim) * .02, requires_grad=True)

indextts/infer.py

@@ -11,6 +11,10 @@ from torch.nn.utils.rnn import pad_sequence
 from omegaconf import OmegaConf
 from tqdm import tqdm
 
+import warnings
+warnings.filterwarnings("ignore", category=FutureWarning)
+warnings.filterwarnings("ignore", category=UserWarning)
+
 from indextts.BigVGAN.models import BigVGAN as Generator
 from indextts.gpt.model import UnifiedVoice
 from indextts.utils.checkpoint import load_checkpoint