deepgeek/CosyVoice
2
0
Fork 0
mirror of https://github.com/FunAudioLLM/CosyVoice.git synced 2026-02-05 18:09:24 +08:00
Code Issues Packages Projects Releases Wiki Activity

fix lint

Browse Source
This commit is contained in:
Yuekai Zhang
2025-09-03 03:45:17 -07:00
parent 633b991290
commit ad257b06e3
4 changed files with 14 additions and 20 deletions
Download Patch File Download Diff File Expand all files Collapse all files

13
runtime/triton_trtllm/model_repo/cosyvoice2/1/model.py
View File

@@ -41,11 +41,11 @@ from transformers import AutoTokenizer
import torchaudio
from matcha.utils.audio import mel_spectrogram
torch.set_num_threads(1)
class TritonPythonModel:
"""Triton Python model for Spark TTS.
@@ -193,7 +193,6 @@ class TritonPythonModel:
return prompt_speech_tokens
def forward_speaker_embedding(self, wav):
"""Forward pass through the speaker embedding component.
@@ -219,7 +218,6 @@ class TritonPythonModel:
return prompt_spk_embedding
def forward_token2wav(
self,
prompt_speech_tokens: torch.Tensor,
@@ -254,7 +252,6 @@ class TritonPythonModel:
inputs_tensor.append(token_offset_tensor)
inputs_tensor.append(finalize_tensor)
# Create and execute inference request
inference_request = pb_utils.InferenceRequest(
model_name='token2wav',
@@ -281,7 +278,6 @@ class TritonPythonModel:
input_ids = torch.cat([input_ids, prompt_speech_tokens], dim=1)
return input_ids
def _extract_speech_feat(self, speech):
speech_feat = mel_spectrogram(
speech,
@@ -299,7 +295,6 @@ class TritonPythonModel:
speech_feat = speech_feat.unsqueeze(dim=0)
return speech_feat
def _llm_gen_thread(self, generated_ids_iter, semantic_token_ids_arr, llm_is_done_flag):
for generated_ids in generated_ids_iter:
generated_ids = generated_ids.tolist()
@@ -338,7 +333,6 @@ class TritonPythonModel:
prompt_speech_feat = speech_feat[:, :2 * token_len].contiguous().half()
prompt_speech_tokens = prompt_speech_tokens[:, :token_len].contiguous()
flow_prompt_speech_token_len = prompt_speech_tokens.shape[-1]
reference_text = pb_utils.get_input_tensor_by_name(request, "reference_text").as_numpy()
@@ -385,7 +379,9 @@ class TritonPythonModel:
this_tts_speech_token = semantic_token_ids_arr[:token_offset + this_token_hop_len + self.flow_pre_lookahead_len]
this_tts_speech_token = torch.tensor(this_tts_speech_token).unsqueeze(dim=0).to(torch.int32).to(self.device)
sub_tts_speech = self.forward_token2wav(prompt_speech_tokens, prompt_speech_feat, prompt_spk_embedding, this_tts_speech_token, request_id, token_offset, False)
sub_tts_speech = self.forward_token2wav(
prompt_speech_tokens, prompt_speech_feat, prompt_spk_embedding,
this_tts_speech_token, request_id, token_offset, False)
audio_tensor = pb_utils.Tensor.from_dlpack("waveform", to_dlpack(sub_tts_speech))
inference_response = pb_utils.InferenceResponse(output_tensors=[audio_tensor])
@@ -413,7 +409,6 @@ class TritonPythonModel:
else:
this_token_hop_len = self.token_hop_len
this_token_hop_len = max(self.token_hop_len, this_token_hop_len)
chunk_index += 1
else:
time.sleep(0.02)

1
runtime/triton_trtllm/model_repo/speaker_embedding/1/model.py
View File

@@ -143,7 +143,6 @@ class TritonPythonModel:
embedding = self._extract_spk_embedding(wav_array)
prompt_spk_embedding_tensor = pb_utils.Tensor.from_dlpack(
"prompt_spk_embedding", to_dlpack(embedding))
inference_response = pb_utils.InferenceResponse(

2
runtime/triton_trtllm/model_repo/token2wav/1/model.py
View File

@@ -49,6 +49,7 @@ logger = logging.getLogger(__name__)
ORIGINAL_VOCAB_SIZE = 151663
torch.set_num_threads(1)
class CosyVoice2:
def __init__(self, model_dir, load_jit=False, load_trt=False, fp16=False, trt_concurrent=1, device='cuda'):
@@ -123,7 +124,6 @@ class CosyVoice2Model:
input_names = ["x", "mask", "mu", "cond"]
return {'min_shape': min_shape, 'opt_shape': opt_shape, 'max_shape': max_shape, 'input_names': input_names}
def token2wav(self, token, prompt_token, prompt_feat, embedding, token_offset, uuid, stream=False, finalize=False, speed=1.0):
with torch.cuda.amp.autocast(self.fp16):
tts_mel, _ = self.flow.inference(token=token.to(self.device),