Upload folder using huggingface_hub

mnt/README

@@ -0,0 +1,8 @@
+PATHS are:
+    
+    model_path = "./mnt/efs/fs1/so-vits-svc-fork-main/logs/44k/female/G_10000.pth"
+    config_path = "./mnt/efs/fs1/so-vits-svc-fork-main/configs_logs/config.json"
+    kmeans_path = "./mnt/efs/fs1/so-vits-svc-fork-main/logs/44k/female/kmeans.pt"
+
+    # Use subprocess to run the 'svc infer' command with the kmeans file
+    cmd = f"svc infer {input_file} -s speaker -r 0.1 -m {model_path} -c {config_path} -k {kmeans_path}"

mnt/efs/fs1/so-vits-svc-fork-main/configs_logs/config.json

@@ -0,0 +1,67 @@
+{
+  "train": {
+    "log_interval": 200,
+    "eval_interval": 800,
+    "seed": 1234,
+    "epochs": 10000,
+    "learning_rate": 0.0001,
+    "betas": [0.8, 0.99],
+    "eps": 1e-9,
+    "batch_size": 16,
+    "fp16_run": false,
+    "bf16_run": false,
+    "lr_decay": 0.999875,
+    "segment_size": 10240,
+    "init_lr_ratio": 1,
+    "warmup_epochs": 0,
+    "c_mel": 45,
+    "c_kl": 1.0,
+    "use_sr": true,
+    "max_speclen": 512,
+    "port": "8001",
+    "keep_ckpts": 3,
+    "num_workers": 4,
+    "log_version": 0,
+    "ckpt_name_by_step": false,
+    "accumulate_grad_batches": 1
+  },
+  "data": {
+    "training_files": "filelists/44k/train.txt",
+    "validation_files": "filelists/44k/val.txt",
+    "max_wav_value": 32768.0,
+    "sampling_rate": 44100,
+    "filter_length": 2048,
+    "hop_length": 512,
+    "win_length": 2048,
+    "n_mel_channels": 80,
+    "mel_fmin": 0.0,
+    "mel_fmax": 22050
+  },
+  "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, 8, 2, 2, 2],
+    "upsample_initial_channel": 512,
+    "upsample_kernel_sizes": [16, 16, 4, 4, 4],
+    "n_layers_q": 3,
+    "use_spectral_norm": false,
+    "gin_channels": 256,
+    "ssl_dim": 256,
+    "n_speakers": 200
+  },
+  "spk": {
+    "Jeanne": 0
+  }
+}

mnt/efs/fs1/so-vits-svc-fork-main/logs/44k/female/D_0.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:635be5c3409aaf3eec4135a1f5a771595683f3a6461ffc5bdea43441e50269a9
+size 187018591

mnt/efs/fs1/so-vits-svc-fork-main/logs/44k/female/D_10000.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:2ec1d02237f77586d6297b9ce294cddc0c5a42163f520cddcc62fefcce8bc8a8
+size 561098249

mnt/efs/fs1/so-vits-svc-fork-main/logs/44k/female/D_9800.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:f533efbe850f9f885da3113d067f9c5a14a290faba0d9b92c8f9cd29a204e1f5
+size 561098249

mnt/efs/fs1/so-vits-svc-fork-main/logs/44k/female/D_9900.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:b78a20be560db15bf0cb79b94b79f34482e81b87a6dba711aef5e5c3b895b3ea
+size 561098249

mnt/efs/fs1/so-vits-svc-fork-main/logs/44k/female/G_0.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:20a327c54e5731bed377bd38404bc32ab98e66a1b2777b0af4cc034d4d6914b0
+size 180628517

mnt/efs/fs1/so-vits-svc-fork-main/logs/44k/female/G_10000.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:4ee4ff04a5938b970ec667b1ce2efdf58178714e873ee7925c942b269add072f
+size 542789469

mnt/efs/fs1/so-vits-svc-fork-main/logs/44k/female/G_9800.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:3ef64f42b06cc32c860f785e74531513cf087cf49f2b0bdf88fb8bfe1eebe301
+size 542789469

mnt/efs/fs1/so-vits-svc-fork-main/logs/44k/female/G_9900.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:3a45e8a9c99d31c9f7ba3aedeec19952821da66da676fb6c171f195754bbb3b7
+size 542789469

mnt/efs/fs1/so-vits-svc-fork-main/logs/44k/female/config.json

@@ -0,0 +1,67 @@
+{
+  "train": {
+    "log_interval": 200,
+    "eval_interval": 800,
+    "seed": 1234,
+    "epochs": 10000,
+    "learning_rate": 0.0001,
+    "betas": [0.8, 0.99],
+    "eps": 1e-9,
+    "batch_size": 16,
+    "fp16_run": false,
+    "bf16_run": false,
+    "lr_decay": 0.999875,
+    "segment_size": 10240,
+    "init_lr_ratio": 1,
+    "warmup_epochs": 0,
+    "c_mel": 45,
+    "c_kl": 1.0,
+    "use_sr": true,
+    "max_speclen": 512,
+    "port": "8001",
+    "keep_ckpts": 3,
+    "num_workers": 4,
+    "log_version": 0,
+    "ckpt_name_by_step": false,
+    "accumulate_grad_batches": 1
+  },
+  "data": {
+    "training_files": "filelists/44k/train.txt",
+    "validation_files": "filelists/44k/val.txt",
+    "max_wav_value": 32768.0,
+    "sampling_rate": 44100,
+    "filter_length": 2048,
+    "hop_length": 512,
+    "win_length": 2048,
+    "n_mel_channels": 80,
+    "mel_fmin": 0.0,
+    "mel_fmax": 22050
+  },
+  "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, 8, 2, 2, 2],
+    "upsample_initial_channel": 512,
+    "upsample_kernel_sizes": [16, 16, 4, 4, 4],
+    "n_layers_q": 3,
+    "use_spectral_norm": false,
+    "gin_channels": 256,
+    "ssl_dim": 256,
+    "n_speakers": 200
+  },
+  "spk": {
+    "Jeanne": 0
+  }
+}

mnt/efs/fs1/so-vits-svc-fork-main/logs/44k/female/kmeans.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:166197acd548c45170e0147b8c3bfb685690fc35bd4cf3f21c885be97640c8df
+size 3076399

mnt/efs/fs1/so-vits-svc-fork-main/logs/44k/lightning_logs/configs.json

@@ -0,0 +1,98 @@
+{
+  "train": {
+    "log_interval": 200,
+    "eval_interval": 800,
+    "seed": 1234,
+    "epochs": 10000,
+    "learning_rate": 0.0001,
+    "betas": [
+      0.8,
+      0.99
+    ],
+    "eps": 1e-09,
+    "batch_size": 16,
+    "fp16_run": false,
+    "bf16_run": false,
+    "lr_decay": 0.999875,
+    "segment_size": 10240,
+    "init_lr_ratio": 1,
+    "warmup_epochs": 0,
+    "c_mel": 45,
+    "c_kl": 1.0,
+    "use_sr": true,
+    "max_speclen": 512,
+    "port": "8001",
+    "keep_ckpts": 3,
+    "num_workers": 4,
+    "log_version": 0,
+    "ckpt_name_by_step": false,
+    "accumulate_grad_batches": 1
+  },
+  "data": {
+    "training_files": "filelists/44k/train.txt",
+    "validation_files": "filelists/44k/val.txt",
+    "max_wav_value": 32768.0,
+    "sampling_rate": 44100,
+    "filter_length": 2048,
+    "hop_length": 512,
+    "win_length": 2048,
+    "n_mel_channels": 80,
+    "mel_fmin": 0.0,
+    "mel_fmax": 22050
+  },
+  "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,
+      8,
+      2,
+      2,
+      2
+    ],
+    "upsample_initial_channel": 512,
+    "upsample_kernel_sizes": [
+      16,
+      16,
+      4,
+      4,
+      4
+    ],
+    "n_layers_q": 3,
+    "use_spectral_norm": false,
+    "gin_channels": 256,
+    "ssl_dim": 256,
+    "n_speakers": 200
+  },
+  "spk": {
+    "Jeanne": 0
+  }
+}

mnt/efs/fs1/so-vits-svc-fork-main/logs/44k/lightning_logs/version_0/events.out.tfevents.1685187272.1a2a580b2318.3938.0

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:ca44a52e1f142f8bb523360dc3377b29cdd64af891078ac6a16e5d10de803338
+size 412700412

mnt/efs/fs1/so-vits-svc-fork-main/logs/44k/lightning_logs/version_0/hparams.yaml

@@ -0,0 +1,83 @@
+data: !!python/object:so_vits_svc_fork.hparams.HParams
+  filter_length: 2048
+  hop_length: 512
+  max_wav_value: 32768.0
+  mel_fmax: 22050
+  mel_fmin: 0.0
+  n_mel_channels: 80
+  sampling_rate: 44100
+  training_files: filelists/44k/train.txt
+  validation_files: filelists/44k/val.txt
+  win_length: 2048
+model: !!python/object:so_vits_svc_fork.hparams.HParams
+  filter_channels: 768
+  gin_channels: 256
+  hidden_channels: 192
+  inter_channels: 192
+  kernel_size: 3
+  n_heads: 2
+  n_layers: 6
+  n_layers_q: 3
+  n_speakers: 200
+  p_dropout: 0.1
+  resblock: '1'
+  resblock_dilation_sizes:
+  - - 1
+    - 3
+    - 5
+  - - 1
+    - 3
+    - 5
+  - - 1
+    - 3
+    - 5
+  resblock_kernel_sizes:
+  - 3
+  - 7
+  - 11
+  ssl_dim: 256
+  upsample_initial_channel: 512
+  upsample_kernel_sizes:
+  - 16
+  - 16
+  - 4
+  - 4
+  - 4
+  upsample_rates:
+  - 8
+  - 8
+  - 2
+  - 2
+  - 2
+  use_spectral_norm: false
+model_dir: logs/44k
+reset_optimizer: false
+spk: !!python/object:so_vits_svc_fork.hparams.HParams
+  Jeanne: 0
+train: !!python/object:so_vits_svc_fork.hparams.HParams
+  accumulate_grad_batches: 1
+  batch_size: 16
+  betas:
+  - 0.8
+  - 0.99
+  bf16_run: false
+  c_kl: 1.0
+  c_mel: 45
+  ckpt_name_by_step: false
+  epochs: 10000
+  eps: 1.0e-09
+  eval_interval: 800
+  fp16_run: false
+  init_lr_ratio: 1
+  keep_ckpts: 3
+  learning_rate: 0.0001
+  log_interval: 200
+  log_version: 0
+  lr_decay: 0.999875
+  max_speclen: 512
+  num_workers: 4
+  port: '8001'
+  seed: 1234
+  segment_size: 10240
+  use_sr: true
+  warmup_epochs: 0

mnt/efs/fs1/so-vits-svc-fork-main/logs/44k_MAY26/D_0.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:635be5c3409aaf3eec4135a1f5a771595683f3a6461ffc5bdea43441e50269a9
+size 187018591

mnt/efs/fs1/so-vits-svc-fork-main/logs/44k_MAY26/D_10000.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:2b849138b2a1a477daeed0bcf9f37a01becaa10750cadc26ce20fd68478271aa
+size 561098249

mnt/efs/fs1/so-vits-svc-fork-main/logs/44k_MAY26/D_9930.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c5fc8736d2462612157c59a372b6dc9d810c71335e81947716f3a0fafacdb7e0
+size 561098249

mnt/efs/fs1/so-vits-svc-fork-main/logs/44k_MAY26/D_9977.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:053baea4862b42480923b16ce4a7ad8981d796fb705ee5bb911387bf4ffef539
+size 561098249

mnt/efs/fs1/so-vits-svc-fork-main/logs/44k_MAY26/G_0.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:20a327c54e5731bed377bd38404bc32ab98e66a1b2777b0af4cc034d4d6914b0
+size 180628517

mnt/efs/fs1/so-vits-svc-fork-main/logs/44k_MAY26/G_10000.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:46e32616166fb7dd7cecfa7185e880d08b3dd9ae5b97ca911bee24a58444fabc
+size 542789469

mnt/efs/fs1/so-vits-svc-fork-main/logs/44k_MAY26/G_9930.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:714e973cec8f49a5ab851c0fe8a70bf19311751ea44d45d4b2eab5e108ee9b7c
+size 542789469

mnt/efs/fs1/so-vits-svc-fork-main/logs/44k_MAY26/G_9977.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:b40f22cd3eeec44f0d05c9484dd21b9c73fb0e4bc9f7d1428871b45858b194f5
+size 542789469

mnt/efs/fs1/so-vits-svc-fork-main/logs/44k_MAY26/config.json

@@ -0,0 +1,67 @@
+{
+  "train": {
+    "log_interval": 200,
+    "eval_interval": 800,
+    "seed": 1234,
+    "epochs": 10000,
+    "learning_rate": 0.0001,
+    "betas": [0.8, 0.99],
+    "eps": 1e-9,
+    "batch_size": 16,
+    "fp16_run": false,
+    "bf16_run": false,
+    "lr_decay": 0.999875,
+    "segment_size": 10240,
+    "init_lr_ratio": 1,
+    "warmup_epochs": 0,
+    "c_mel": 45,
+    "c_kl": 1.0,
+    "use_sr": true,
+    "max_speclen": 512,
+    "port": "8001",
+    "keep_ckpts": 3,
+    "num_workers": 4,
+    "log_version": 0,
+    "ckpt_name_by_step": false,
+    "accumulate_grad_batches": 1
+  },
+  "data": {
+    "training_files": "filelists/44k/train.txt",
+    "validation_files": "filelists/44k/val.txt",
+    "max_wav_value": 32768.0,
+    "sampling_rate": 44100,
+    "filter_length": 2048,
+    "hop_length": 512,
+    "win_length": 2048,
+    "n_mel_channels": 80,
+    "mel_fmin": 0.0,
+    "mel_fmax": 22050
+  },
+  "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, 8, 2, 2, 2],
+    "upsample_initial_channel": 512,
+    "upsample_kernel_sizes": [16, 16, 4, 4, 4],
+    "n_layers_q": 3,
+    "use_spectral_norm": false,
+    "gin_channels": 256,
+    "ssl_dim": 256,
+    "n_speakers": 200
+  },
+  "spk": {
+    "Jeanne": 0
+  }
+}

mnt/efs/fs1/so-vits-svc-fork-main/logs/44k_MAY26/kmeans.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:f641ba93c287d45ef48ebe80942b4a2d76413478a9d438c3637e98bf0683e230
+size 3077423

mnt/efs/fs1/so-vits-svc-fork-main/logs/44k_MAY26/lightning_logs/version_0/events.out.tfevents.1684895655.34b5025d974f.7410.0

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:6071639195f7dea96bb967ef7bbc36224801f51087f39d078de2a1f863d6f402
+size 746318639

mnt/efs/fs1/so-vits-svc-fork-main/logs/44k_MAY26/lightning_logs/version_0/hparams.yaml

@@ -0,0 +1,83 @@
+data: !!python/object:so_vits_svc_fork.hparams.HParams
+  filter_length: 2048
+  hop_length: 512
+  max_wav_value: 32768.0
+  mel_fmax: 22050
+  mel_fmin: 0.0
+  n_mel_channels: 80
+  sampling_rate: 44100
+  training_files: filelists/44k/train.txt
+  validation_files: filelists/44k/val.txt
+  win_length: 2048
+model: !!python/object:so_vits_svc_fork.hparams.HParams
+  filter_channels: 768
+  gin_channels: 256
+  hidden_channels: 192
+  inter_channels: 192
+  kernel_size: 3
+  n_heads: 2
+  n_layers: 6
+  n_layers_q: 3
+  n_speakers: 200
+  p_dropout: 0.1
+  resblock: '1'
+  resblock_dilation_sizes:
+  - - 1
+    - 3
+    - 5
+  - - 1
+    - 3
+    - 5
+  - - 1
+    - 3
+    - 5
+  resblock_kernel_sizes:
+  - 3
+  - 7
+  - 11
+  ssl_dim: 256
+  upsample_initial_channel: 512
+  upsample_kernel_sizes:
+  - 16
+  - 16
+  - 4
+  - 4
+  - 4
+  upsample_rates:
+  - 8
+  - 8
+  - 2
+  - 2
+  - 2
+  use_spectral_norm: false
+model_dir: logs/44k
+reset_optimizer: false
+spk: !!python/object:so_vits_svc_fork.hparams.HParams
+  Jeanne: 0
+train: !!python/object:so_vits_svc_fork.hparams.HParams
+  accumulate_grad_batches: 1
+  batch_size: 16
+  betas:
+  - 0.8
+  - 0.99
+  bf16_run: false
+  c_kl: 1.0
+  c_mel: 45
+  ckpt_name_by_step: false
+  epochs: 10000
+  eps: 1.0e-09
+  eval_interval: 800
+  fp16_run: false
+  init_lr_ratio: 1
+  keep_ckpts: 3
+  learning_rate: 0.0001
+  log_interval: 200
+  log_version: 0
+  lr_decay: 0.999875
+  max_speclen: 512
+  num_workers: 4
+  port: '8001'
+  seed: 1234
+  segment_size: 10240
+  use_sr: true
+  warmup_epochs: 0