{ "train": { "log_interval": 200, "eval_interval": 1000, "seed": 42, "epochs": 1000, "learning_rate": 0.00007, "betas": [ 0.8, 0.99 ], "eps": 1e-09, "batch_size": 6, "bf16_run": false, "fp16_run": false, "lr_decay": 0.99996, "segment_size": 16384, "init_lr_ratio": 1, "warmup_epochs": 0, "c_mel": 45, "c_kl": 1.0, "c_commit": 1.0, "skip_optimizer": true, "freeze_ZH_bert": false, "freeze_JP_bert": false, "freeze_EN_bert": false, "freeze_emo": false }, "data": { "training_files": "filelists/train.list", "validation_files": "filelists/val.list", "max_wav_value": 32768.0, "sampling_rate": 44100, "filter_length": 2048, "hop_length": 512, "win_length": 2048, "n_mel_channels": 128, "mel_fmin": 0.0, "mel_fmax": null, "add_blank": true, "n_speakers": 192, "cleaned_text": true, "spk2id": { "派蒙_ZH": 0, "纳西妲_ZH": 1, "娜维娅_ZH": 2, "荒泷一斗_ZH": 3, "凯亚_ZH": 4, "林尼_ZH": 5, "温迪_ZH": 6, "阿贝多_ZH": 7, "芙宁娜_ZH": 8, "钟离_ZH": 9, "赛诺_ZH": 10, "那维莱特_ZH": 11, "提纳里_ZH": 12, "枫原万叶_ZH": 13, "艾尔海森_ZH": 14, "八重神子_ZH": 15, "宵宫_ZH": 16, "卡维_ZH": 17, "迪希雅_ZH": 18, "莱依拉_ZH": 19, "诺艾尔_ZH": 20, "托马_ZH": 21, "莫娜_ZH": 22, "凝光_ZH": 23, "神里绫华_ZH": 24, "北斗_ZH": 25, "莱欧斯利_ZH": 26, "柯莱_ZH": 27, "迪奥娜_ZH": 28, "可莉_ZH": 29, "丽莎_ZH": 30, "琳妮特_ZH": 31, "五郎_ZH": 32, "雷电将军_ZH": 33, "芭芭拉_ZH": 34, "珊瑚宫心海_ZH": 35, "鹿野院平藏_ZH": 36, "魈_ZH": 37, "达达利亚_ZH": 38, "琴_ZH": 39, "胡桃_ZH": 40, "砂糖_ZH": 41, "安柏_ZH": 42, "重云_ZH": 43, "夜兰_ZH": 44, "班尼特_ZH": 45, "珐露珊_ZH": 46, "妮露_ZH": 47, "辛焱_ZH": 48, "香菱_ZH": 49, "迪卢克_ZH": 50, "刻晴_ZH": 51, "烟绯_ZH": 52, "久岐忍_ZH": 53, "早柚_ZH": 54, "云堇_ZH": 55, "夏洛蒂_ZH": 56, "夏沃蕾_ZH": 57, "优菈_ZH": 58, "克洛琳德_ZH": 59, "神里绫人_ZH": 60, "甘雨_ZH": 61, "流浪者_ZH": 62, "行秋_ZH": 63, "千织_ZH": 64, "戴因斯雷布_ZH": 65, "希格雯_ZH": 66, "阿蕾奇诺_ZH": 67, "闲云_ZH": 68, "白术_ZH": 69, "菲谢尔_ZH": 70, "荧_ZH": 71, "空_ZH": 72, "申鹤_ZH": 73, "九条裟罗_ZH": 74, "菲米尼_ZH": 75, "雷泽_ZH": 76, "嘉明_ZH": 77, "多莉_ZH": 78, "凯瑟琳_ZH": 79, "迪娜泽黛_ZH": 80, "绮良良_ZH": 81, "坎蒂丝_ZH": 82, "罗莎莉亚_ZH": 83, "米卡_ZH": 84, "萍姥姥_ZH": 85, "赛索斯_ZH": 86, "留云借风真君_ZH": 87, "埃德_ZH": 88, "爱贝尔_ZH": 89, "瑶瑶_ZH": 90, "伊迪娅_ZH": 91, "七七_ZH": 92, "式大将_ZH": 93, "奥兹_ZH": 94, "德沃沙克_ZH": 95, "泽维尔_ZH": 96, "哲平_ZH": 97, "大肉丸_ZH": 98, "浮游水蕈兽·元素生命_ZH": 99, "蒂玛乌斯_ZH": 100, "塞琉斯_ZH": 101, "欧菲妮_ZH": 102, "昆钧_ZH": 103, "主持人_ZH": 104, "言笑_ZH": 105, "迈勒斯_ZH": 106, "杜拉夫_ZH": 107, "拉赫曼_ZH": 108, "旁白_ZH": 109, "伊利亚斯_ZH": 110, "爱德琳_ZH": 111, "居勒什_ZH": 112, "埃洛伊_ZH": 113, "远黛_ZH": 114, "弗洛莱恩_ZH": 115, "柊千里_ZH": 116, "塞塔蕾_ZH": 117, "海芭夏_ZH": 118, "回声海螺_ZH": 119, "阿扎尔_ZH": 120, "博士_ZH": 121, "天叔_ZH": 122, "祖莉亚·德斯特雷_ZH": 123, "三月七_ZH": 124, "瓦尔特_ZH": 125, "丹恒_ZH": 126, "砂金_ZH": 127, "希儿_ZH": 128, "姬子_ZH": 129, "流萤_ZH": 130, "穹_ZH": 131, "星_ZH": 132, "希露瓦_ZH": 133, "娜塔莎_ZH": 134, "黄泉_ZH": 135, "黑天鹅_ZH": 136, "佩拉_ZH": 137, "布洛妮娅_ZH": 138, "虎克_ZH": 139, "素裳_ZH": 140, "克拉拉_ZH": 141, "符玄_ZH": 142, "白露_ZH": 143, "景元_ZH": 144, "杰帕德_ZH": 145, "藿藿_ZH": 146, "星期日_ZH": 147, "桑博_ZH": 148, "卡芙卡_ZH": 149, "艾丝妲_ZH": 150, "托帕_ZH": 151, "真理医生_ZH": 152, "桂乃芬_ZH": 153, "加拉赫_ZH": 154, "米沙_ZH": 155, "知更鸟_ZH": 156, "彦卿_ZH": 157, "玲可_ZH": 158, "波提欧_ZH": 159, "黑塔_ZH": 160, "驭空_ZH": 161, "停云_ZH": 162, "镜流_ZH": 163, "银枝_ZH": 164, "银狼_ZH": 165, "卢卡_ZH": 166, "帕姆_ZH": 167, "罗刹_ZH": 168, "阮•梅_ZH": 169, "翡翠_ZH": 170, "青雀_ZH": 171, "阿兰_ZH": 172, "浮烟_ZH": 173, "螺丝咕姆_ZH": 174, "花火_ZH": 175, "史瓦罗_ZH": 176, "明曦_ZH": 177, "寒鸦_ZH": 178, "雪衣_ZH": 179, "乔瓦尼_ZH": 180, "伦纳德_ZH": 181, "公输师傅_ZH": 182, "晴霓_ZH": 183, "奥列格_ZH": 184, "丹枢_ZH": 185, "刃_ZH": 186, "尾巴_ZH": 187, "可可利亚_ZH": 188, "青镞_ZH": 189, "半夏_ZH": 190, "梦主_ZH": 191 } }, "model": { "use_spk_conditioned_encoder": true, "use_noise_scaled_mas": true, "use_mel_posterior_encoder": false, "use_duration_discriminator": false, "use_wavlm_discriminator": true, "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": [ 8, 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": 256, "slm": { "model": "./slm/wavlm-base-plus", "sr": 16000, "hidden": 768, "nlayers": 13, "initial_channel": 64 } }, "version": "2.4" }