Duplicate from Plachta/Seed-VC

Co-authored-by: ElderFrog <Plachta@users.noreply.huggingface.co>

.gitattributes

@@ -0,0 +1,35 @@
+*.7z filter=lfs diff=lfs merge=lfs -text
+*.arrow filter=lfs diff=lfs merge=lfs -text
+*.bin filter=lfs diff=lfs merge=lfs -text
+*.bz2 filter=lfs diff=lfs merge=lfs -text
+*.ckpt filter=lfs diff=lfs merge=lfs -text
+*.ftz filter=lfs diff=lfs merge=lfs -text
+*.gz filter=lfs diff=lfs merge=lfs -text
+*.h5 filter=lfs diff=lfs merge=lfs -text
+*.joblib filter=lfs diff=lfs merge=lfs -text
+*.lfs.* filter=lfs diff=lfs merge=lfs -text
+*.mlmodel filter=lfs diff=lfs merge=lfs -text
+*.model filter=lfs diff=lfs merge=lfs -text
+*.msgpack filter=lfs diff=lfs merge=lfs -text
+*.npy filter=lfs diff=lfs merge=lfs -text
+*.npz filter=lfs diff=lfs merge=lfs -text
+*.onnx filter=lfs diff=lfs merge=lfs -text
+*.ot filter=lfs diff=lfs merge=lfs -text
+*.parquet filter=lfs diff=lfs merge=lfs -text
+*.pb filter=lfs diff=lfs merge=lfs -text
+*.pickle filter=lfs diff=lfs merge=lfs -text
+*.pkl filter=lfs diff=lfs merge=lfs -text
+*.pt filter=lfs diff=lfs merge=lfs -text
+*.pth filter=lfs diff=lfs merge=lfs -text
+*.rar filter=lfs diff=lfs merge=lfs -text
+*.safetensors filter=lfs diff=lfs merge=lfs -text
+saved_model/**/* filter=lfs diff=lfs merge=lfs -text
+*.tar.* filter=lfs diff=lfs merge=lfs -text
+*.tar filter=lfs diff=lfs merge=lfs -text
+*.tflite filter=lfs diff=lfs merge=lfs -text
+*.tgz filter=lfs diff=lfs merge=lfs -text
+*.wasm filter=lfs diff=lfs merge=lfs -text
+*.xz filter=lfs diff=lfs merge=lfs -text
+*.zip filter=lfs diff=lfs merge=lfs -text
+*.zst filter=lfs diff=lfs merge=lfs -text
+*tfevents* filter=lfs diff=lfs merge=lfs -text

DiT_seed_v2_uvit_whisper_base_f0_44k_bigvgan_pruned_ema.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:4145da6704ca0950c2ef6312c1cd0c15c370b13613faed2b4a9d9d71bb4ed036
+size 820863652

DiT_seed_v2_uvit_whisper_base_f0_44k_bigvgan_pruned_ft_ema.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:883d19be9fe9c4bb2dce30d545cd77e5ed2b6326c7ed1c4d9a7ce8e555b163f9
+size 820864573

DiT_seed_v2_uvit_whisper_base_f0_44k_bigvgan_pruned_ft_ema_v2.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:42aef93ffe65857c840d270252fa040f7ba04514945ec460f3ac1ac2a96de684
+size 820865494

DiT_seed_v2_uvit_whisper_small_wavenet_bigvgan_pruned.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:8ec8841b20bb46df9f7e8e570a6946a4b87b940133c7f0e778487ff33841f720
+size 440312082

DiT_uvit_tat_xlsr_ema.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c853ea578b409f625f961bcb15d5cff1f8ef9a75f3209ec21d9b7c73ab422e88
+size 142112203

README.md

@@ -0,0 +1,4 @@
+---
+license: gpl-3.0
+pipeline_tag: audio-to-audio
+---

config_dit_mel_seed_uvit_whisper_base_f0_44k.yml

@@ -0,0 +1,98 @@
+log_dir: "./runs/run_dit_mel_seed_uvit_whisper_base_f0_44k"
+save_freq: 1
+log_interval: 10
+save_interval: 1000
+device: "cuda"
+epochs: 1000 # number of epochs for first stage training (pre-training)
+batch_size: 1
+batch_length: 100 # maximum duration of audio in a batch (in seconds)
+max_len: 80 # maximum number of frames
+pretrained_model: ""
+pretrained_encoder: ""
+load_only_params: False # set to true if do not want to load epoch numbers and optimizer parameters
+
+preprocess_params:
+  sr: 44100
+  spect_params:
+    n_fft: 2048
+    win_length: 2048
+    hop_length: 512
+    n_mels: 128
+    fmin: 0
+    fmax: "None"
+
+model_params:
+  dit_type: "DiT" # uDiT or DiT
+  reg_loss_type: "l1" # l1 or l2
+
+  timbre_shifter:
+    se_db_path: "./modules/openvoice/checkpoints_v2/converter/se_db.pt"
+    ckpt_path: './modules/openvoice/checkpoints_v2/converter'
+
+  vocoder:
+    type: "bigvgan"
+    name: "nvidia/bigvgan_v2_44khz_128band_512x"
+
+  speech_tokenizer:
+    type: 'whisper'
+    name: "openai/whisper-small"
+
+  style_encoder:
+    dim: 192
+    campplus_path: "campplus_cn_common.bin"
+
+  DAC:
+    encoder_dim: 64
+    encoder_rates: [2, 5, 5, 6]
+    decoder_dim: 1536
+    decoder_rates: [ 6, 5, 5, 2 ]
+    sr: 24000
+
+  length_regulator:
+    channels: 768
+    is_discrete: false
+    in_channels: 768
+    content_codebook_size: 2048
+    sampling_ratios: [1, 1, 1, 1]
+    vector_quantize: false
+    n_codebooks: 1
+    quantizer_dropout: 0.0
+    f0_condition: true
+    n_f0_bins: 256
+
+  DiT:
+    hidden_dim: 768
+    num_heads: 12
+    depth: 17
+    class_dropout_prob: 0.1
+    block_size: 8192
+    in_channels: 128
+    style_condition: true
+    final_layer_type: 'mlp'
+    target: 'mel' # mel or codec
+    content_dim: 768
+    content_codebook_size: 1024
+    content_type: 'discrete'
+    f0_condition: true
+    n_f0_bins: 256
+    content_codebooks: 1
+    is_causal: false
+    long_skip_connection: false
+    zero_prompt_speech_token: false # for prompt component, do not input corresponding speech token
+    time_as_token: false
+    style_as_token: false
+    uvit_skip_connection: true
+    add_resblock_in_transformer: false
+
+  wavenet:
+    hidden_dim: 768
+    num_layers: 8
+    kernel_size: 5
+    dilation_rate: 1
+    p_dropout: 0.2
+    style_condition: true
+
+loss_params:
+  base_lr: 0.0001
+  lambda_mel: 45
+  lambda_kl: 1.0

config_dit_mel_seed_uvit_whisper_small_wavenet.yml

@@ -0,0 +1,91 @@
+log_dir: "./runs/run_dit_mel_seed_uvit_whisper_small_wavenet"
+save_freq: 1
+log_interval: 10
+save_interval: 1000
+device: "cuda"
+epochs: 1000 # number of epochs for first stage training (pre-training)
+batch_size: 2
+batch_length: 100 # maximum duration of audio in a batch (in seconds)
+max_len: 80 # maximum number of frames
+pretrained_model: ""
+pretrained_encoder: "./temp_ckpt.pth"
+load_only_params: False # set to true if do not want to load epoch numbers and optimizer parameters
+
+preprocess_params:
+  sr: 22050
+  spect_params:
+    n_fft: 1024
+    win_length: 1024
+    hop_length: 256
+    n_mels: 80
+    fmin: 0
+    fmax: "None"
+
+model_params:
+  dit_type: "DiT" # uDiT or DiT
+  reg_loss_type: "l1" # l1 or l2
+
+  timbre_shifter:
+    se_db_path: "./modules/openvoice/checkpoints_v2/converter/se_db.pt"
+    ckpt_path: './modules/openvoice/checkpoints_v2/converter'
+
+  speech_tokenizer:
+    type: 'whisper'
+    name: "openai/whisper-small"
+
+  style_encoder:
+    dim: 192
+    campplus_path: "campplus_cn_common.bin"
+
+  vocoder:
+    type: "bigvgan"
+    name: "nvidia/bigvgan_v2_22khz_80band_256x"
+
+  length_regulator:
+    channels: 512
+    is_discrete: false
+    in_channels: 768
+    content_codebook_size: 2048
+    sampling_ratios: [1, 1, 1, 1]
+    vector_quantize: false
+    n_codebooks: 1
+    quantizer_dropout: 0.0
+    f0_condition: false
+    n_f0_bins: 512
+
+  DiT:
+    hidden_dim: 512
+    num_heads: 8
+    depth: 13
+    class_dropout_prob: 0.1
+    block_size: 8192
+    in_channels: 80
+    style_condition: true
+    final_layer_type: 'wavenet'
+    target: 'mel' # mel or codec
+    content_dim: 512
+    content_codebook_size: 1024
+    content_type: 'discrete'
+    f0_condition: false
+    n_f0_bins: 512
+    content_codebooks: 1
+    is_causal: false
+    long_skip_connection: true
+    zero_prompt_speech_token: false # for prompt component, do not input corresponding speech token
+    time_as_token: false
+    style_as_token: false
+    uvit_skip_connection: true
+    add_resblock_in_transformer: false
+
+  wavenet:
+    hidden_dim: 512
+    num_layers: 8
+    kernel_size: 5
+    dilation_rate: 1
+    p_dropout: 0.2
+    style_condition: true
+
+loss_params:
+  base_lr: 0.0001
+  lambda_mel: 45
+  lambda_kl: 1.0

config_dit_mel_seed_uvit_xlsr_tiny.yml

@@ -0,0 +1,82 @@
+log_dir: "runs/run_mel_seed_uvit_xlsr_tiny"
+save_freq: 1
+log_interval: 10
+save_interval: 500
+device: "cuda"
+epochs: 1000 # number of epochs for first stage training (pre-training)
+batch_size: 2
+batch_length: 100 # maximum duration of audio in a batch (in seconds)
+max_len: 80 # maximum number of frames
+pretrained_model: "DiT_uvit_tat_xlsr_ema.pth"
+pretrained_encoder: ""
+load_only_params: False # set to true if do not want to load epoch numbers and optimizer parameters
+
+preprocess_params:
+  sr: 22050
+  spect_params:
+    n_fft: 1024
+    win_length: 1024
+    hop_length: 256
+    n_mels: 80
+    fmin: 0
+    fmax: 8000
+
+model_params:
+  dit_type: "DiT" # uDiT or DiT
+  reg_loss_type: "l1" # l1 or l2
+  diffusion_type: "flow"
+
+  timbre_shifter:
+    se_db_path: "./modules/openvoice/checkpoints_v2/converter/se_db.pt"
+    ckpt_path: './modules/openvoice/checkpoints_v2/converter'
+
+  vocoder:
+    type: "hifigan"
+
+  speech_tokenizer:
+    type: 'xlsr'
+    output_layer: 12
+    name: 'facebook/wav2vec2-xls-r-300m'
+
+  style_encoder:
+    dim: 192
+    campplus_path: "campplus_cn_common.bin"
+
+  length_regulator:
+    channels: 384
+    is_discrete: false
+    in_channels: 1024
+    content_codebook_size: 1024
+    sampling_ratios: [1, 1, 1, 1]
+    vector_quantize: false
+    n_codebooks: 2
+    quantizer_dropout: 0.0
+    f0_condition: false
+    n_f0_bins: 512
+
+  DiT:
+    hidden_dim: 384
+    num_heads: 6
+    depth: 9
+    class_dropout_prob: 0.1
+    block_size: 8192
+    in_channels: 80
+    style_condition: true
+    final_layer_type: 'mlp'
+    target: 'mel' # mel or betavae
+    content_dim: 384
+    content_codebook_size: 1024
+    content_type: 'discrete'
+    f0_condition: false
+    n_f0_bins: 512
+    content_codebooks: 1
+    is_causal: false
+    long_skip_connection: false
+    zero_prompt_speech_token: false # for prompt component, do not input corresponding speech token
+    time_as_token: true
+    style_as_token: true
+    uvit_skip_connection: true
+    add_resblock_in_transformer: false
+
+loss_params:
+  base_lr: 0.0001

config_dit_mel_seed_wavenet.yml

@@ -0,0 +1,79 @@
+log_dir: "./runs/run_dit_mel_seed"
+save_freq: 1
+log_interval: 10
+save_interval: 1000
+device: "cuda"
+epochs: 1000 # number of epochs for first stage training (pre-training)
+batch_size: 4
+batch_length: 100 # maximum duration of audio in a batch (in seconds)
+max_len: 80 # maximum number of frames
+pretrained_model: ""
+pretrained_encoder: ""
+load_only_params: False # set to true if do not want to load epoch numbers and optimizer parameters
+
+F0_path: "modules/JDC/bst.t7"
+
+preprocess_params:
+  sr: 22050
+  spect_params:
+    n_fft: 1024
+    win_length: 1024
+    hop_length: 256
+    n_mels: 80
+
+model_params:
+  dit_type: "DiT" # uDiT or DiT
+  reg_loss_type: "l2" # l1 or l2
+
+  speech_tokenizer:
+    path: "speech_tokenizer_v1.onnx"
+
+  style_encoder:
+    dim: 192
+    campplus_path: "campplus_cn_common.bin"
+
+  DAC:
+    encoder_dim: 64
+    encoder_rates: [2, 5, 5, 6]
+    decoder_dim: 1536
+    decoder_rates: [ 6, 5, 5, 2 ]
+    sr: 24000
+
+  length_regulator:
+    channels: 768
+    is_discrete: true
+    content_codebook_size: 4096
+    in_frame_rate: 50
+    out_frame_rate: 80
+    sampling_ratios: [1, 1, 1, 1]
+
+  DiT:
+    hidden_dim: 768
+    num_heads: 12
+    depth: 12
+    class_dropout_prob: 0.1
+    block_size: 4096
+    in_channels: 80
+    style_condition: true
+    final_layer_type: 'wavenet'
+    target: 'mel' # mel or codec
+    content_dim: 768
+    content_codebook_size: 1024
+    content_type: 'discrete'
+    f0_condition: false
+    n_f0_bins: 512
+    content_codebooks: 1
+    is_causal: false
+    long_skip_connection: true
+    zero_prompt_speech_token: false # for prompt component, do not input corresponding speech token
+
+  wavenet:
+    hidden_dim: 768
+    num_layers: 8
+    kernel_size: 5
+    dilation_rate: 1
+    p_dropout: 0.2
+    style_condition: true
+
+loss_params:
+  base_lr: 0.0001

hifigan.yml

@@ -0,0 +1,25 @@
+hift:
+    in_channels: 80
+    base_channels: 512
+    nb_harmonics: 8
+    sampling_rate: 22050
+    nsf_alpha: 0.1
+    nsf_sigma: 0.003
+    nsf_voiced_threshold: 10
+    upsample_rates: [8, 8]
+    upsample_kernel_sizes: [16, 16]
+    istft_params:
+        n_fft: 16
+        hop_len: 4
+    resblock_kernel_sizes: [3, 7, 11]
+    resblock_dilation_sizes: [[1, 3, 5], [1, 3, 5], [1, 3, 5]]
+    source_resblock_kernel_sizes: [7, 11]
+    source_resblock_dilation_sizes: [[1, 3, 5], [1, 3, 5]]
+    lrelu_slope: 0.1
+    audio_limit: 0.99
+f0_predictor:
+    num_class: 1
+    in_channels: 80
+    cond_channels: 512
+
+pretrained_model_path: "hift.pt"

hift.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:91e679b6ca1eff71187ffb4f3ab0444935594cdcc20a9bd12afad111ef8d6012
+size 81896716

se_db.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:a76e1ad4fac609cfa1a5198bf789427c5e7822d656fab1ffed6cd80fc0d381b7
+size 102405286

v2/ar_base.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:42e2afabf97e5d856580af1f81a9705423b4fb21bb9475cece9925b62355cc5c
+size 358846893

v2/cfm_small.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:73ae18615dc8cb01710086629973198226dedc63ebeaab0aee17863d767b3474
+size 353349694