Closed5

[MMVC] バッチサイズ(batch size)の確認

Srgr0Srgr0

MMVCについて

構成

ターゲット:ずんだもん
ユーザー音声サンプル:ITAコーパス 100文

train_config.json

注記:batch_sizeのみ変更して実行しています

{
  "train": {
    "eval_interval": 1000,
    "best": true,
    "backup": {
      "interval": 2000,
      "g_only": true,
      "mean_of_num_eval": 10
    },
    "seed": 1234,
    "learning_rate": 0.0002,
    "betas": [
      0.8,
      0.99
    ],
    "eps": 1e-09,
    "batch_size": 60,
    "fp16_run": true,
    "lr_decay": 0.999875,
    "segment_size": 8192,
    "init_lr_ratio": 1,
    "warmup_epochs": 0,
    "c_mel": 45,
    "c_kl": 1.0
  },
  "data": {
    "training_files": "filelists/train_config_textful.txt",
    "validation_files": "filelists/train_config_textful_val.txt",
    "training_files_notext": "filelists/train_config_textless.txt",
    "validation_files_notext": "filelists/train_config_val_textless.txt",
    "text_cleaners": [
      "japanese_cleaners"
    ],
    "max_wav_value": 32768.0,
    "sampling_rate": 24000,
    "filter_length": 512,
    "hop_length": 128,
    "win_length": 512,
    "n_mel_channels": 80,
    "mel_fmin": 0.0,
    "mel_fmax": null,
    "add_blank": true,
    "n_speakers": 110,
    "cleaned_text": false
  },
  "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": [
      8,
      4,
      2,
      2
    ],
    "upsample_initial_channel": 512,
    "upsample_kernel_sizes": [
      16,
      16,
      8,
      8
    ],
    "n_layers_q": 3,
    "use_spectral_norm": false,
    "n_flow": 8,
    "gin_channels": 256,
    "use_mel_train": false
  },
  "others": {
    "os_type": "linux",
    "input_filename": "dataset/textful/00_myvoice/wav/emotion002.wav",
    "source_id": 107,
    "target_id": 100
  },
  "augmentation": {
    "enable": true,
    "gain_p": 0.5,
    "min_gain_in_db": -10,
    "max_gain_in_db": 10,
    "time_stretch_p": 0.5,
    "min_rate": 0.75,
    "max_rate": 1.25,
    "pitch_shift_p": 0.0,
    "min_semitones": -4.0,
    "max_semitones": 4.0,
    "add_gaussian_noise_p": 0.0,
    "min_amplitude": 0.001,
    "max_amplitude": 0.04,
    "frequency_mask_p": 0.0
  }
}

Template

Colab
Env: Colab
Instance: Standard GPU, Standard GPU(highmem), Premium GPU, Premium GPU(highmem)
CPU: 2vCPU
RAM: xxxGB
GPU: T4, A100 etc.
GPU RAM: xxxGB
result_batch_size: xxx
result_cpu_usage: xxx
result_ram_usage: xxx
result_gpu_ram_usage: xxx

!nvidia-smi
xxxxxxxxxxxxxxxxxxxx

!train_ms.py
xxxxxxxxxxxxxxxxxxxx
GCP(Colab VM), GCP, AWS etc.
Env: GCP(Colab VM), GCP, AWS etc.
Instance: n1-highmem-2, g4dn.xlarge etc.
OS: Container-Optimized OS(Colab VM), Ubuntu, Debian etc.
CPU: xxx
RAM: xxxGB
GPU: T4, V100, A100 etc.
GPU RAM: xxxGB
result_batch_size: xxx
result_cpu_usage: xxx
result_ram_usage: xxx
result_gpu_ram_usage: xxx

!nvidia-smi
xxxxxxxxxxxxxxxxxxxx

!train_ms.py
xxxxxxxxxxxxxxxxxxxx
Local
Env: Local
Instance: Model name (if you know)
OS: Ubuntu(WSL2), Ubuntu etc.
CPU: xxx
RAM: xxxGB
GPU: GTX 1080, RTX 3090 etc.
GPU RAM: xxxGB
result_batch_size: xxx
result_cpu_usage: xxx
result_ram_usage: xxx
result_gpu_ram_usage: xxx

!nvidia-smi
xxxxxxxxxxxxxxxxxxxx

!train_ms.py
xxxxxxxxxxxxxxxxxxxx

Srgr0Srgr0

Env: GCP(Colab VM)
Instance: n1-highmem-2
OS: Container-Optimized OS(Colab VM)
CPU: 2 vCPU
RAM: 13 GB
GPU: T4
GPU RAM: 16 GB GDDR6@ 320 GB/s
result_batch_size: 18 (20はepoch4でout of memory)
result_cpu_usage: 62-65%
result_ram_usage: no-data
result_gpu_ram_usage: no-data

nvidia-smi
!nvidia-smi
+-----------------------------------------------------------------------------+
| NVIDIA-SMI 460.32.03    Driver Version: 460.32.03    CUDA Version: 11.2     |
|-------------------------------+----------------------+----------------------+
| GPU  Name        Persistence-M| Bus-Id        Disp.A | Volatile Uncorr. ECC |
| Fan  Temp  Perf  Pwr:Usage/Cap|         Memory-Usage | GPU-Util  Compute M. |
|                               |                      |               MIG M. |
|===============================+======================+======================|
|   0  Tesla T4            Off  | 00000000:00:04.0 Off |                    0 |
| N/A   51C    P8     9W /  70W |      0MiB / 15109MiB |      0%      Default |
|                               |                      |                  N/A |
+-------------------------------+----------------------+----------------------+
                                                                               
+-----------------------------------------------------------------------------+
| Processes:                                                                  |
|  GPU   GI   CI        PID   Type   Process name                  GPU Memory |
|        ID   ID                                                   Usage      |
|=============================================================================|
|  No running processes found                                                 |
+-----------------------------------------------------------------------------+
train_ms.py
!train_ms.py
[INFO] {'train': {'eval_interval': 1000, 'best': True, 'backup': {'interval': 2000, 'g_only': True, 'mean_of_num_eval': 10}, 'seed': 1234, 'learning_rate': 0.0002, 'betas': [0.8, 0.99], 'eps': 1e-09, 'batch_size': 18, 'fp16_run': True, 'lr_decay': 0.999875, 'segment_size': 8192, 'init_lr_ratio': 1, 'warmup_epochs': 0, 'c_mel': 45, 'c_kl': 1.0}, 'data': {'training_files': 'filelists/train_config_textful.txt', 'validation_files': 'filelists/train_config_textful_val.txt', 'training_files_notext': 'filelists/train_config_textless.txt', 'validation_files_notext': 'filelists/train_config_val_textless.txt', 'text_cleaners': ['japanese_cleaners'], 'max_wav_value': 32768.0, 'sampling_rate': 24000, 'filter_length': 512, 'hop_length': 128, 'win_length': 512, 'n_mel_channels': 80, 'mel_fmin': 0.0, 'mel_fmax': None, 'add_blank': True, 'n_speakers': 110, 'cleaned_text': False}, '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': [8, 4, 2, 2], 'upsample_initial_channel': 512, 'upsample_kernel_sizes': [16, 16, 8, 8], 'n_layers_q': 3, 'use_spectral_norm': False, 'n_flow': 8, 'gin_channels': 256, 'use_mel_train': False}, 'others': {'os_type': 'linux', 'input_filename': 'dataset/textful/00_myvoice/wav/emotion002.wav', 'source_id': 107, 'target_id': 100}, 'augmentation': {'enable': True, 'gain_p': 0.5, 'min_gain_in_db': -10, 'max_gain_in_db': 10, 'time_stretch_p': 0.5, 'min_rate': 0.75, 'max_rate': 1.25, 'pitch_shift_p': 0.0, 'min_semitones': -4.0, 'max_semitones': 4.0, 'add_gaussian_noise_p': 0.0, 'min_amplitude': 0.001, 'max_amplitude': 0.04, 'frequency_mask_p': 0.0}, 'fine_flag': True, 'fine_model_g': 'fine_model/G_180000.pth', 'fine_model_d': 'fine_model/D_180000.pth', 'model_dir': './logs/20220306_24000', 'best_log_path': './logs/20220306_24000/best.log', 'best_loss_mel': 9999}
[WARNING] /content/MMVC_Trainer-main is not a git repository, therefore hash value comparison will be ignored.
100% 472/472 [00:00<00:00, 224660.86it/s]
100% 53/53 [00:00<00:00, 142865.11it/s]
[INFO] FineTuning : True
[INFO] Load model : fine_model/G_180000.pth
[INFO] Load model : fine_model/D_180000.pth
[INFO] Loaded checkpoint 'fine_model/G_180000.pth' (iteration 637)
[INFO] Loaded checkpoint 'fine_model/D_180000.pth' (iteration 637)
Epoch 1: 100% 31/31 [01:42<00:00,  3.31s/it]
Epoch 2: 100% 31/31 [01:24<00:00,  2.72s/it]
Epoch 3: 100% 31/31 [01:21<00:00,  2.61s/it]
Epoch 4: 100% 31/31 [01:22<00:00,  2.67s/it]
Epoch 5: 100% 31/31 [01:24<00:00,  2.72s/it]
Epoch 6: 100% 31/31 [01:18<00:00,  2.52s/it]
Epoch 7: 100% 31/31 [01:15<00:00,  2.45s/it]
Epoch 8: 100% 31/31 [01:23<00:00,  2.69s/it]
Srgr0Srgr0

Env: GCP(Colab VM)
Instance: n1-highmem-2
OS: Container-Optimized OS(Colab VM)
CPU: 2 vCPU
RAM: 13 GB
GPU: P100
GPU RAM: 16 GB HBM2@ 732 GB/s
result_batch_size: 20 (22はepoch3でout of memory)
result_cpu_usage: no-data
result_ram_usage: no-data
result_gpu_ram_usage: no-data

nvidia-smi
!nvidia-smi
+-----------------------------------------------------------------------------+
| NVIDIA-SMI 460.32.03    Driver Version: 460.32.03    CUDA Version: 11.2     |
|-------------------------------+----------------------+----------------------+
| GPU  Name        Persistence-M| Bus-Id        Disp.A | Volatile Uncorr. ECC |
| Fan  Temp  Perf  Pwr:Usage/Cap|         Memory-Usage | GPU-Util  Compute M. |
|                               |                      |               MIG M. |
|===============================+======================+======================|
|   0  Tesla P100-PCIE...  Off  | 00000000:00:04.0 Off |                    0 |
| N/A   33C    P0    26W / 250W |      0MiB / 16280MiB |      0%      Default |
|                               |                      |                  N/A |
+-------------------------------+----------------------+----------------------+
                                                                               
+-----------------------------------------------------------------------------+
| Processes:                                                                  |
|  GPU   GI   CI        PID   Type   Process name                  GPU Memory |
|        ID   ID                                                   Usage      |
|=============================================================================|
|  No running processes found                                                 |
+-----------------------------------------------------------------------------+
train_ms.py
!train_ms.py
[INFO] {'train': {'eval_interval': 1000, 'best': True, 'backup': {'interval': 2000, 'g_only': True, 'mean_of_num_eval': 10}, 'seed': 1234, 'learning_rate': 0.0002, 'betas': [0.8, 0.99], 'eps': 1e-09, 'batch_size': 20, 'fp16_run': True, 'lr_decay': 0.999875, 'segment_size': 8192, 'init_lr_ratio': 1, 'warmup_epochs': 0, 'c_mel': 45, 'c_kl': 1.0}, 'data': {'training_files': 'filelists/train_config_textful.txt', 'validation_files': 'filelists/train_config_textful_val.txt', 'training_files_notext': 'filelists/train_config_textless.txt', 'validation_files_notext': 'filelists/train_config_val_textless.txt', 'text_cleaners': ['japanese_cleaners'], 'max_wav_value': 32768.0, 'sampling_rate': 24000, 'filter_length': 512, 'hop_length': 128, 'win_length': 512, 'n_mel_channels': 80, 'mel_fmin': 0.0, 'mel_fmax': None, 'add_blank': True, 'n_speakers': 110, 'cleaned_text': False}, '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': [8, 4, 2, 2], 'upsample_initial_channel': 512, 'upsample_kernel_sizes': [16, 16, 8, 8], 'n_layers_q': 3, 'use_spectral_norm': False, 'n_flow': 8, 'gin_channels': 256, 'use_mel_train': False}, 'others': {'os_type': 'linux', 'input_filename': 'dataset/textful/00_myvoice/wav/emotion002.wav', 'source_id': 107, 'target_id': 100}, 'augmentation': {'enable': True, 'gain_p': 0.5, 'min_gain_in_db': -10, 'max_gain_in_db': 10, 'time_stretch_p': 0.5, 'min_rate': 0.75, 'max_rate': 1.25, 'pitch_shift_p': 0.0, 'min_semitones': -4.0, 'max_semitones': 4.0, 'add_gaussian_noise_p': 0.0, 'min_amplitude': 0.001, 'max_amplitude': 0.04, 'frequency_mask_p': 0.0}, 'fine_flag': True, 'fine_model_g': 'fine_model/G_180000.pth', 'fine_model_d': 'fine_model/D_180000.pth', 'model_dir': './logs/20220306_24000', 'best_log_path': './logs/20220306_24000/best.log', 'best_loss_mel': 9999}
[WARNING] /content/MMVC_Trainer-main is not a git repository, therefore hash value comparison will be ignored.
100% 472/472 [00:00<00:00, 230440.17it/s]
100% 53/53 [00:00<00:00, 202826.74it/s]
[INFO] FineTuning : True
[INFO] Load model : fine_model/G_180000.pth
[INFO] Load model : fine_model/D_180000.pth
[INFO] Loaded checkpoint 'fine_model/G_180000.pth' (iteration 637)
[INFO] Loaded checkpoint 'fine_model/D_180000.pth' (iteration 637)
Epoch 1: 100% 28/28 [01:26<00:00,  3.09s/it]
Epoch 2: 100% 28/28 [01:10<00:00,  2.53s/it]
Epoch 3: 100% 28/28 [01:09<00:00,  2.49s/it]
Epoch 4: 100% 28/28 [01:10<00:00,  2.53s/it]
Epoch 5: 100% 28/28 [01:10<00:00,  2.53s/it]
Epoch 6: 100% 28/28 [01:10<00:00,  2.51s/it]
Epoch 7: 100% 28/28 [01:09<00:00,  2.48s/it]
Epoch 8: 100% 28/28 [01:10<00:00,  2.52s/it]
Srgr0Srgr0

Env: GCP(Colab VM)
Instance: n1-highmem-2
OS: Container-Optimized OS(Colab VM)
CPU: 2 vCPU
RAM: 13 GB
GPU: V100
GPU RAM: 16 GB HBM2@ 900 GB/s
result_batch_size: 18 (20はepoch4でout of memory)
result_cpu_usage: no-data
result_ram_usage: no-data
result_gpu_ram_usage: no-data

nvidia-smi
!nvidia-smi
+-----------------------------------------------------------------------------+
| NVIDIA-SMI 460.32.03    Driver Version: 460.32.03    CUDA Version: 11.2     |
|-------------------------------+----------------------+----------------------+
| GPU  Name        Persistence-M| Bus-Id        Disp.A | Volatile Uncorr. ECC |
| Fan  Temp  Perf  Pwr:Usage/Cap|         Memory-Usage | GPU-Util  Compute M. |
|                               |                      |               MIG M. |
|===============================+======================+======================|
|   0  Tesla V100-SXM2...  Off  | 00000000:00:04.0 Off |                    0 |
| N/A   32C    P0    24W / 300W |      0MiB / 16160MiB |      0%      Default |
|                               |                      |                  N/A |
+-------------------------------+----------------------+----------------------+
                                                                               
+-----------------------------------------------------------------------------+
| Processes:                                                                  |
|  GPU   GI   CI        PID   Type   Process name                  GPU Memory |
|        ID   ID                                                   Usage      |
|=============================================================================|
|  No running processes found                                                 |
+-----------------------------------------------------------------------------+
train_ms.py
!train_ms.py
[INFO] {'train': {'eval_interval': 1000, 'best': True, 'backup': {'interval': 2000, 'g_only': True, 'mean_of_num_eval': 10}, 'seed': 1234, 'learning_rate': 0.0002, 'betas': [0.8, 0.99], 'eps': 1e-09, 'batch_size': 18, 'fp16_run': True, 'lr_decay': 0.999875, 'segment_size': 8192, 'init_lr_ratio': 1, 'warmup_epochs': 0, 'c_mel': 45, 'c_kl': 1.0}, 'data': {'training_files': 'filelists/train_config_textful.txt', 'validation_files': 'filelists/train_config_textful_val.txt', 'training_files_notext': 'filelists/train_config_textless.txt', 'validation_files_notext': 'filelists/train_config_val_textless.txt', 'text_cleaners': ['japanese_cleaners'], 'max_wav_value': 32768.0, 'sampling_rate': 24000, 'filter_length': 512, 'hop_length': 128, 'win_length': 512, 'n_mel_channels': 80, 'mel_fmin': 0.0, 'mel_fmax': None, 'add_blank': True, 'n_speakers': 110, 'cleaned_text': False}, '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': [8, 4, 2, 2], 'upsample_initial_channel': 512, 'upsample_kernel_sizes': [16, 16, 8, 8], 'n_layers_q': 3, 'use_spectral_norm': False, 'n_flow': 8, 'gin_channels': 256, 'use_mel_train': False}, 'others': {'os_type': 'linux', 'input_filename': 'dataset/textful/00_myvoice/wav/emotion002.wav', 'source_id': 107, 'target_id': 100}, 'augmentation': {'enable': True, 'gain_p': 0.5, 'min_gain_in_db': -10, 'max_gain_in_db': 10, 'time_stretch_p': 0.5, 'min_rate': 0.75, 'max_rate': 1.25, 'pitch_shift_p': 0.0, 'min_semitones': -4.0, 'max_semitones': 4.0, 'add_gaussian_noise_p': 0.0, 'min_amplitude': 0.001, 'max_amplitude': 0.04, 'frequency_mask_p': 0.0}, 'fine_flag': True, 'fine_model_g': 'fine_model/G_180000.pth', 'fine_model_d': 'fine_model/D_180000.pth', 'model_dir': './logs/20220306_24000', 'best_log_path': './logs/20220306_24000/best.log', 'best_loss_mel': 9999}
[WARNING] /content/MMVC_Trainer-main is not a git repository, therefore hash value comparison will be ignored.
100% 472/472 [00:00<00:00, 209759.64it/s]
100% 53/53 [00:00<00:00, 208534.81it/s]
[INFO] FineTuning : True
[INFO] Load model : fine_model/G_180000.pth
[INFO] Load model : fine_model/D_180000.pth
[INFO] Loaded checkpoint 'fine_model/G_180000.pth' (iteration 637)
[INFO] Loaded checkpoint 'fine_model/D_180000.pth' (iteration 637)
Epoch 1: 100% 31/31 [00:59<00:00,  1.93s/it]
Epoch 2: 100% 31/31 [00:44<00:00,  1.45s/it]
Epoch 3: 100% 31/31 [00:44<00:00,  1.45s/it]
Epoch 4: 100% 31/31 [00:44<00:00,  1.44s/it]
Epoch 5: 100% 31/31 [00:46<00:00,  1.49s/it]
Epoch 6: 100% 31/31 [00:43<00:00,  1.39s/it]
Epoch 7: 100% 31/31 [00:42<00:00,  1.38s/it]
Epoch 8: 100% 31/31 [00:44<00:00,  1.43s/it]
Srgr0Srgr0

Env: GCP(Colab VM)
Instance: a2-highgpu-1g
OS: Container-Optimized OS(Colab VM)
CPU: 12 vCPU
RAM: 85 GB
GPU: A100 40GB
GPU RAM: 40 GB HBM2@ 1.6 TB/s
result_batch_size: 60 (66はepoch1でout of memory)
result_cpu_usage: 14-17%
result_ram_usage: 8.0-8.8GB
result_gpu_ram_usage: no-data

nvidia-smi
!nvidia-smi
+-----------------------------------------------------------------------------+
| NVIDIA-SMI 460.32.03    Driver Version: 460.32.03    CUDA Version: 11.2     |
|-------------------------------+----------------------+----------------------+
| GPU  Name        Persistence-M| Bus-Id        Disp.A | Volatile Uncorr. ECC |
| Fan  Temp  Perf  Pwr:Usage/Cap|         Memory-Usage | GPU-Util  Compute M. |
|                               |                      |               MIG M. |
|===============================+======================+======================|
|   0  A100-SXM4-40GB      Off  | 00000000:00:04.0 Off |                    0 |
| N/A   30C    P0    44W / 400W |      0MiB / 40536MiB |      0%      Default |
|                               |                      |             Disabled |
+-------------------------------+----------------------+----------------------+
                                                                               
+-----------------------------------------------------------------------------+
| Processes:                                                                  |
|  GPU   GI   CI        PID   Type   Process name                  GPU Memory |
|        ID   ID                                                   Usage      |
|=============================================================================|
|  No running processes found                                                 |
+-----------------------------------------------------------------------------+
train_ms.py
!train_ms.py
[INFO] {'train': {'eval_interval': 1000, 'best': True, 'backup': {'interval': 2000, 'g_only': True, 'mean_of_num_eval': 10}, 'seed': 1234, 'learning_rate': 0.0002, 'betas': [0.8, 0.99], 'eps': 1e-09, 'batch_size': 60, 'fp16_run': True, 'lr_decay': 0.999875, 'segment_size': 8192, 'init_lr_ratio': 1, 'warmup_epochs': 0, 'c_mel': 45, 'c_kl': 1.0}, 'data': {'training_files': 'filelists/train_config_textful.txt', 'validation_files': 'filelists/train_config_textful_val.txt', 'training_files_notext': 'filelists/train_config_textless.txt', 'validation_files_notext': 'filelists/train_config_val_textless.txt', 'text_cleaners': ['japanese_cleaners'], 'max_wav_value': 32768.0, 'sampling_rate': 24000, 'filter_length': 512, 'hop_length': 128, 'win_length': 512, 'n_mel_channels': 80, 'mel_fmin': 0.0, 'mel_fmax': None, 'add_blank': True, 'n_speakers': 110, 'cleaned_text': False}, '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': [8, 4, 2, 2], 'upsample_initial_channel': 512, 'upsample_kernel_sizes': [16, 16, 8, 8], 'n_layers_q': 3, 'use_spectral_norm': False, 'n_flow': 8, 'gin_channels': 256, 'use_mel_train': False}, 'others': {'os_type': 'linux', 'input_filename': 'dataset/textful/00_myvoice/wav/emotion002.wav', 'source_id': 107, 'target_id': 100}, 'augmentation': {'enable': True, 'gain_p': 0.5, 'min_gain_in_db': -10, 'max_gain_in_db': 10, 'time_stretch_p': 0.5, 'min_rate': 0.75, 'max_rate': 1.25, 'pitch_shift_p': 0.0, 'min_semitones': -4.0, 'max_semitones': 4.0, 'add_gaussian_noise_p': 0.0, 'min_amplitude': 0.001, 'max_amplitude': 0.04, 'frequency_mask_p': 0.0}, 'fine_flag': True, 'fine_model_g': 'fine_model/G_180000.pth', 'fine_model_d': 'fine_model/D_180000.pth', 'model_dir': './logs/20220306_24000', 'best_log_path': './logs/20220306_24000/best.log', 'best_loss_mel': 9999}
[WARNING] /content/MMVC_Trainer-main is not a git repository, therefore hash value comparison will be ignored.
2023-01-09 11:27:52.454744: I tensorflow/core/util/util.cc:169] oneDNN custom operations are on. You may see slightly different numerical results due to floating-point round-off errors from different computation orders. To turn them off, set the environment variable `TF_ENABLE_ONEDNN_OPTS=0`.
100% 472/472 [00:00<00:00, 219382.92it/s]
100% 53/53 [00:00<00:00, 204318.12it/s]
[INFO] FineTuning : True
[INFO] Load model : fine_model/G_180000.pth
[INFO] Load model : fine_model/D_180000.pth
[INFO] Loaded checkpoint 'fine_model/G_180000.pth' (iteration 637)
[INFO] Loaded checkpoint 'fine_model/D_180000.pth' (iteration 637)
Epoch 1: 100% 13/13 [01:02<00:00,  4.78s/it]
Epoch 2: 100% 13/13 [00:50<00:00,  3.91s/it]
Epoch 3: 100% 13/13 [00:45<00:00,  3.51s/it]
Epoch 4: 100% 13/13 [00:45<00:00,  3.53s/it]
Epoch 5: 100% 13/13 [00:49<00:00,  3.79s/it]
Epoch 6: 100% 13/13 [00:48<00:00,  3.69s/it]
Epoch 7: 100% 13/13 [00:47<00:00,  3.64s/it]
Epoch 8: 100% 13/13 [00:47<00:00,  3.68s/it]
このスクラップは2023/01/15にクローズされました