GPT-SoVITS模型|男声口语化声音模型|男声口语化配音模型|

声音特点适用领域  ：磁性温和，峥气十足，角色，娱乐，影视，资讯，百科

模型配音效果

鉴于GPT-SoVITS模型自回归特性，即其配音情绪高度依赖于所提供的参考音频，特此说明：本视频所展示的配音情绪仅为采用某一特定参考音频时的效果示例，并不全面反映GPT-SoVITS模型能够生成的全部情绪范围及最终配音质量的上限。模型的最终表现将随着不同参考音频的输入而展现出多样化。

模型下载

训练日志

2025-07-10 15:32:06,611	peiyin.me_男声口语_理性_新闻	INFO	{'train': {'log_interval': 100, 'eval_interval': 500, 'seed': 1234, 'epochs': 20, 'learning_rate': 0.0001, 'betas': [0.8, 0.99], 'eps': 1e-09, 'batch_size': 11, 'fp16_run': True, 'lr_decay': 0.999875, 'segment_size': 20480, 'init_lr_ratio': 1, 'warmup_epochs': 0, 'c_mel': 45, 'c_kl': 1.0, 'text_low_lr_rate': 0.4, 'pretrained_s2G': 'GPT_SoVITS/pretrained_models/gsv-v2final-pretrained/s2G2333k.pth', 'pretrained_s2D': 'GPT_SoVITS/pretrained_models/gsv-v2final-pretrained/s2D2333k.pth', 'if_save_latest': True, 'if_save_every_weights': True, 'save_every_epoch': 20, 'gpu_numbers': '0'}, 'data': {'max_wav_value': 32768.0, 'sampling_rate': 32000, 'filter_length': 2048, 'hop_length': 640, 'win_length': 2048, 'n_mel_channels': 128, 'mel_fmin': 0.0, 'mel_fmax': None, 'add_blank': True, 'n_speakers': 300, 'cleaned_text': True, 'exp_dir': 'logs/peiyin.me_男声口语_理性_新闻'}, 'model': {'inter_channels': 192, 'hidden_channels': 192, 'filter_channels': 768, 'n_heads': 2, 'n_layers': 6, 'kernel_size': 3, 'p_dropout': 0.1, 'resblock': '1', 'resblock_kernel_sizes': [3, 7, 11], 'resblock_dilation_sizes': [[1, 3, 5], [1, 3, 5], [1, 3, 5]], 'upsample_rates': [10, 8, 2, 2, 2], 'upsample_initial_channel': 512, 'upsample_kernel_sizes': [16, 16, 8, 2, 2], 'n_layers_q': 3, 'use_spectral_norm': False, 'gin_channels': 512, 'semantic_frame_rate': '25hz', 'freeze_quantizer': True, 'version': 'v2'}, 's2_ckpt_dir': 'logs/peiyin.me_男声口语_理性_新闻', 'content_module': 'cnhubert', 'save_weight_dir': 'SoVITS_weights_v2', 'name': 'peiyin.me_男声口语_理性_新闻', 'version': 'v2', 'pretrain': None, 'resume_step': None}
2025-07-10 15:32:07,936	peiyin.me_男声口语_理性_新闻	INFO	loaded pretrained GPT_SoVITS/pretrained_models/gsv-v2final-pretrained/s2G2333k.pth
2025-07-10 15:32:08,169	peiyin.me_男声口语_理性_新闻	INFO	loaded pretrained GPT_SoVITS/pretrained_models/gsv-v2final-pretrained/s2D2333k.pth
2025-07-10 15:32:42,813	peiyin.me_男声口语_理性_新闻	INFO	Train Epoch: 1 [0%]
2025-07-10 15:32:42,814	peiyin.me_男声口语_理性_新闻	INFO	[2.8384013175964355, 1.9210718870162964, 9.815083503723145, 25.536624908447266, 0.0, 2.400508403778076, 0, 9.99875e-05]
2025-07-10 15:32:54,776	peiyin.me_男声口语_理性_新闻	INFO	====> Epoch: 1
2025-07-10 15:33:07,130	peiyin.me_男声口语_理性_新闻	INFO	====> Epoch: 2
2025-07-10 15:33:18,937	peiyin.me_男声口语_理性_新闻	INFO	====> Epoch: 3
2025-07-10 15:33:30,572	peiyin.me_男声口语_理性_新闻	INFO	====> Epoch: 4
2025-07-10 15:33:42,781	peiyin.me_男声口语_理性_新闻	INFO	====> Epoch: 5
2025-07-10 15:33:54,286	peiyin.me_男声口语_理性_新闻	INFO	====> Epoch: 6
2025-07-10 15:34:03,386	peiyin.me_男声口语_理性_新闻	INFO	Train Epoch: 7 [67%]
2025-07-10 15:34:03,386	peiyin.me_男声口语_理性_新闻	INFO	[2.9000988006591797, 2.3289952278137207, 10.491010665893555, 25.393634796142578, 0.0, 2.004318952560425, 100, 9.991253280566489e-05]
2025-07-10 15:34:06,340	peiyin.me_男声口语_理性_新闻	INFO	====> Epoch: 7
2025-07-10 15:34:18,423	peiyin.me_男声口语_理性_新闻	INFO	====> Epoch: 8
2025-07-10 15:34:30,040	peiyin.me_男声口语_理性_新闻	INFO	====> Epoch: 9
2025-07-10 15:34:41,563	peiyin.me_男声口语_理性_新闻	INFO	====> Epoch: 10
2025-07-10 15:34:53,194	peiyin.me_男声口语_理性_新闻	INFO	====> Epoch: 11
2025-07-10 15:35:04,755	peiyin.me_男声口语_理性_新闻	INFO	====> Epoch: 12
2025-07-10 15:35:16,616	peiyin.me_男声口语_理性_新闻	INFO	====> Epoch: 13
2025-07-10 15:35:22,486	peiyin.me_男声口语_理性_新闻	INFO	Train Epoch: 14 [33%]
2025-07-10 15:35:22,486	peiyin.me_男声口语_理性_新闻	INFO	[2.547726631164551, 2.507960557937622, 10.415836334228516, 24.877918243408203, 0.0, 1.9603184461593628, 200, 9.982514211643064e-05]
2025-07-10 15:35:28,712	peiyin.me_男声口语_理性_新闻	INFO	====> Epoch: 14
2025-07-10 15:35:40,316	peiyin.me_男声口语_理性_新闻	INFO	====> Epoch: 15
2025-07-10 15:35:52,264	peiyin.me_男声口语_理性_新闻	INFO	====> Epoch: 16
2025-07-10 15:36:04,901	peiyin.me_男声口语_理性_新闻	INFO	====> Epoch: 17
2025-07-10 15:36:16,678	peiyin.me_男声口语_理性_新闻	INFO	====> Epoch: 18
2025-07-10 15:36:28,451	peiyin.me_男声口语_理性_新闻	INFO	====> Epoch: 19
2025-07-10 15:36:40,451	peiyin.me_男声口语_理性_新闻	INFO	Saving model and optimizer state at iteration 20 to logs/peiyin.me_男声口语_理性_新闻/logs_s2\G_233333333333.pth
2025-07-10 15:36:41,353	peiyin.me_男声口语_理性_新闻	INFO	Saving model and optimizer state at iteration 20 to logs/peiyin.me_男声口语_理性_新闻/logs_s2\D_233333333333.pth
2025-07-10 15:36:43,096	peiyin.me_男声口语_理性_新闻	INFO	saving ckpt peiyin.me_男声口语_理性_新闻_e20:Success.
2025-07-10 15:36:43,096	peiyin.me_男声口语_理性_新闻	INFO	====> Epoch: 20

如何使用配音模型

1，GPT-SoVITS模型本地部署（适合有显卡的用户或者好的CPU）

https://wanlei.cc/6268/.html

2，GPT-SoVITS模型云端部署

https://wanlei.cc/99/.html

客服微信 xiaoming1870