GPT-SOVITS模型|新三国曹操声音配音模型|陈建斌曹操AI配音模型

10天前更新

140

角色介绍

在电视剧《三国》中饰演曹操一角，其表演特点鲜明：展现了曹操的足智多谋、军事才能和霸气侧漏的一面，同时细腻地刻画了曹操复杂多变的情感世界。尽管有观众对其文人气质的表现提出不同看法，但整体上，陈建斌成功地塑造了一个生动、立体的曹操形象，为观众呈现了一个既威严又智慧的历史人物。

配音模型效果

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。

训练日志

2024-08-03 11:54:02,394	曹操	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': 20, '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/s2G488k.pth', 'pretrained_s2D': 'GPT_SoVITS/pretrained_models/s2D488k.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/曹操'}, '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}, 's2_ckpt_dir': 'logs/曹操', 'content_module': 'cnhubert', 'save_weight_dir': 'SoVITS_weights', 'name': '曹操', 'pretrain': None, 'resume_step': None}
2024-08-03 11:54:03,991	曹操	INFO	loaded pretrained GPT_SoVITS/pretrained_models/s2G488k.pth
2024-08-03 11:54:04,211	曹操	INFO	loaded pretrained GPT_SoVITS/pretrained_models/s2D488k.pth
2024-08-03 11:54:39,351	曹操	INFO	Train Epoch: 1 [0%]
2024-08-03 11:54:39,351	曹操	INFO	[2.884860038757324, 1.7871466875076294, 15.621310234069824, 22.641021728515625, 0.0, 3.2314882278442383, 0, 9.99875e-05]
2024-08-03 11:54:52,371	曹操	INFO	====> Epoch: 1
2024-08-03 11:55:06,539	曹操	INFO	====> Epoch: 2
2024-08-03 11:55:19,868	曹操	INFO	====> Epoch: 3
2024-08-03 11:55:33,059	曹操	INFO	====> Epoch: 4
2024-08-03 11:55:46,273	曹操	INFO	====> Epoch: 5
2024-08-03 11:55:59,632	曹操	INFO	====> Epoch: 6
2024-08-03 11:56:13,488	曹操	INFO	====> Epoch: 7
2024-08-03 11:56:24,525	曹操	INFO	Train Epoch: 8 [69%]
2024-08-03 11:56:24,525	曹操	INFO	[2.326380729675293, 2.8958120346069336, 13.948054313659668, 22.318513870239258, 0.0, 2.129373550415039, 100, 9.990004373906418e-05]
2024-08-03 11:56:27,412	曹操	INFO	====> Epoch: 8
2024-08-03 11:56:40,503	曹操	INFO	====> Epoch: 9
2024-08-03 11:56:53,636	曹操	INFO	====> Epoch: 10
2024-08-03 11:57:06,705	曹操	INFO	====> Epoch: 11
2024-08-03 11:57:19,804	曹操	INFO	====> Epoch: 12
2024-08-03 11:57:32,981	曹操	INFO	====> Epoch: 13
2024-08-03 11:57:46,278	曹操	INFO	====> Epoch: 14
2024-08-03 11:57:59,386	曹操	INFO	====> Epoch: 15
2024-08-03 11:58:07,347	曹操	INFO	Train Epoch: 16 [38%]
2024-08-03 11:58:07,348	曹操	INFO	[2.334193468093872, 2.556783437728882, 15.669257164001465, 21.29042625427246, 0.0, 2.0957415103912354, 200, 9.980018739066937e-05]
2024-08-03 11:58:13,423	曹操	INFO	====> Epoch: 16
2024-08-03 11:58:27,143	曹操	INFO	====> Epoch: 17
2024-08-03 11:58:40,505	曹操	INFO	====> Epoch: 18
2024-08-03 11:58:53,566	曹操	INFO	====> Epoch: 19
2024-08-03 11:59:06,667	曹操	INFO	Saving model and optimizer state at iteration 20 to logs/曹操/logs_s2\G_233333333333.pth
2024-08-03 11:59:07,570	曹操	INFO	Saving model and optimizer state at iteration 20 to logs/曹操/logs_s2\D_233333333333.pth
2024-08-03 11:59:09,374	曹操	INFO	saving ckpt 曹操_e20:Success.
2024-08-03 11:59:09,374	曹操	INFO	====> Epoch: 20