Upload folder using huggingface_hub

mdx23c/mdx23c_sfx_jasper.ckpt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:527f0a98e106fc4e2ad939a8c09d4f59966aa13272c31a530b36598f81c0edd3
+size 1344401799

mdx23c/mdx23c_sfx_jasper_config.yaml

@@ -0,0 +1,54 @@
+audio:
+  chunk_size: 261120
+  dim_f: 4096
+  dim_t: 256
+  hop_length: 1024
+  n_fft: 8192
+  num_channels: 2
+  sample_rate: 44100
+  min_mean_abs: 0.001
+
+model:
+  act: gelu
+  bottleneck_factor: 4
+  growth: 128
+  norm: InstanceNorm
+  num_blocks_per_scale: 2
+  num_channels: 128
+  num_scales: 5
+  num_subbands: 4
+  scale:
+  - 2
+  - 2
+
+training:
+  batch_size: 6
+  gradient_accumulation_steps: 1
+  grad_clip: 0
+  instruments:
+  - foreground
+  - background
+  lr: 9.0e-05
+  patience: 2
+  reduce_factor: 0.95
+  target_instrument: null
+  num_epochs: 1000
+  num_steps: 1000
+  augmentation: false # enable augmentations by audiomentations and pedalboard
+  augmentation_type: simple1
+  use_mp3_compress: false # Deprecated
+  augmentation_mix: true # Mix several stems of the same type with some probability
+  augmentation_loudness: true # randomly change loudness of each stem
+  augmentation_loudness_type: 1 # Type 1 or 2
+  augmentation_loudness_min: 0.5
+  augmentation_loudness_max: 1.5
+  q: 0.95
+  coarse_loss_clip: true
+  ema_momentum: 0.999
+  optimizer: adam
+  other_fix: true # it's needed for checking on multisong dataset if other is actually instrumental
+
+inference:
+  batch_size: 1
+  dim_t: 256
+  num_overlap: 4

mel_band_roformer/mbr_bgm_jasper.ckpt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:49248d3e106e95f351b4a51942b87c00b6c3d3f6c0da11f6cf4d86b2d8ba55f9
+size 2739398173

mel_band_roformer/mbr_bgm_jasper_config.yaml

@@ -0,0 +1,72 @@
+audio:
+  chunk_size: 352800
+  dim_f: 1024
+  dim_t: 256
+  hop_length: 441
+  n_fft: 2048
+  num_channels: 2
+  sample_rate: 44100
+  min_mean_abs: 0.000
+
+model:
+  dim: 384
+  depth: 6
+  stereo: true
+  num_stems: 1
+  time_transformer_depth: 1
+  freq_transformer_depth: 1
+  num_bands: 60
+  dim_head: 64
+  heads: 8
+  attn_dropout: 0
+  ff_dropout: 0
+  flash_attn: True
+  dim_freqs_in: 1025
+  sample_rate: 44100  # needed for mel filter bank from librosa
+  stft_n_fft: 2048
+  stft_hop_length: 441
+  stft_win_length: 2048
+  stft_normalized: False
+  mask_estimator_depth: 2
+  multi_stft_resolution_loss_weight: 1.0
+  multi_stft_resolutions_window_sizes: !!python/tuple
+  - 4096
+  - 2048
+  - 1024
+  - 512
+  - 256
+  multi_stft_hop_size: 147
+  multi_stft_normalized: False
+
+training:
+  batch_size: 2
+  gradient_accumulation_steps: 1
+  grad_clip: 0
+  instruments:
+  - vocals
+  - other
+  lr: 1.0e-05
+  patience: 2
+  reduce_factor: 0.95
+  target_instrument: vocals
+  num_epochs: 1000
+  num_steps: 1000
+  augmentation: false # enable augmentations by audiomentations and pedalboard
+  augmentation_type: null
+  use_mp3_compress: false # Deprecated
+  augmentation_mix: false # Mix several stems of the same type with some probability
+  augmentation_loudness: false # randomly change loudness of each stem
+  augmentation_loudness_type: 1 # Type 1 or 2
+  augmentation_loudness_min: 0
+  augmentation_loudness_max: 0
+  q: 0.95
+  coarse_loss_clip: false
+  ema_momentum: 0.999
+  optimizer: adam
+  other_fix: true # it's needed for checking on multisong dataset if other is actually instrumental
+  use_amp: true # enable or disable usage of mixed precision (float16) - usually it must be true
+
+inference:
+  batch_size: 4
+  dim_t: 256
+  num_overlap: 2

mel_band_roformer/mbr_guitar_chencfd.ckpt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:1a81a87b46eed215c051579ac5cff3b139bfdfb9fa3bf4666df059ef87404d14
+size 276726667

mel_band_roformer/mbr_guitar_chencfd_config.yaml

@@ -0,0 +1,74 @@
+audio:
+  chunk_size: 131584
+  dim_f: 1024
+  dim_t: 256
+  hop_length: 512
+  n_fft: 2048
+  num_channels: 2
+  sample_rate: 44100
+  min_mean_abs: 0.001
+
+model:
+  dim: 192
+  depth: 8
+  stereo: true
+  num_stems: 1
+  time_transformer_depth: 1
+  freq_transformer_depth: 1
+  linear_transformer_depth: 0
+  num_bands: 60
+  dim_head: 64
+  heads: 8
+  attn_dropout: 0.1
+  ff_dropout: 0.1
+  flash_attn: True
+  dim_freqs_in: 1025
+  sample_rate: 44100 # needed for mel filter bank from librosa
+  stft_n_fft: 2048
+  stft_hop_length: 512
+  stft_win_length: 2048
+  stft_normalized: False
+  mask_estimator_depth: 2
+  multi_stft_resolution_loss_weight: 1.0
+  multi_stft_resolutions_window_sizes: !!python/tuple
+    - 4096
+    - 2048
+    - 1024
+    - 512
+    - 256
+  multi_stft_hop_size: 147
+  multi_stft_normalized: False
+  mlp_expansion_factor: 4 # Probably too big (requires a lot of memory for weights)
+  use_torch_checkpoint: False  # it allows to greatly reduce GPU memory consumption during training (not fully tested)
+  skip_connection: False # Enable skip connection between transformer blocks - can solve problem with gradients and probably faster training
+
+training:
+  batch_size: 7
+  gradient_accumulation_steps: 1
+  grad_clip: 0
+  instruments:
+    - guitar
+    - others
+  lr: 5.0e-05
+  patience: 2
+  reduce_factor: 0.95
+  target_instrument: guitar
+  num_epochs: 1000
+  num_steps: 1000
+  q: 0.95
+  coarse_loss_clip: true
+  ema_momentum: 0.999
+  optimizer: adam
+  other_fix: false # it's needed for checking on multisong dataset if other is actually instrumental
+  use_amp: true # enable or disable usage of mixed precision (float16) - usually it must be true
+
+augmentations:
+  enable: true # enable or disable all augmentations (to fast disable if needed)
+  loudness: true # randomly change loudness of each stem on the range (loudness_min; loudness_max)
+  loudness_min: 0.5
+  loudness_max: 1.5
+
+inference:
+  batch_size: 1
+  dim_t: 256
+  num_overlap: 4

scnet_masked/scnet_masked_choirsep_exp.ckpt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:1f4982b7e6facf41a81fbf00244cd20597fbbe2939b24731fcfe417651b62a5c
+size 42522921

scnet_masked/scnet_masked_choirsep_exp_config.yaml

@@ -0,0 +1,152 @@
+audio:
+  chunk_size: 88200 # 44100 * 11
+  num_channels: 2
+  sample_rate: 44100
+  min_mean_abs: 0.000
+
+model:
+  sources:
+    - soprano
+    - alto
+    - tenor
+    - bass
+  audio_channels: 2
+  dims:
+    - 4
+    - 32
+    - 64
+    - 128
+  nfft: 4096
+  hop_size: 1024
+  win_size: 4096
+  normalized: True
+  band_SR:
+    - 0.175
+    - 0.392
+    - 0.433
+  band_stride:
+    - 1
+    - 4
+    - 16
+  band_kernel:
+    - 3
+    - 4
+    - 16
+  conv_depths:
+    - 3
+    - 2
+    - 1
+  compress: 4
+  conv_kernel: 3
+  num_dplayer: 6
+  expand: 1
+
+loss_multistft:
+  fft_sizes:
+  - 1024
+  - 2048
+  - 4096
+  hop_sizes:
+  - 512
+  - 1024
+  - 2048
+  win_lengths:
+  - 1024
+  - 2048
+  - 4096
+  window: "hann_window"
+  scale: "mel"
+  n_bins: 128
+  sample_rate: 44100
+  perceptual_weighting: true
+  w_sc: 1.0
+  w_log_mag: 1.0
+  w_lin_mag: 0.0
+  w_phs: 0.0
+  mag_distance: "L1"
+
+
+training:
+  batch_size: 6
+  gradient_accumulation_steps: 1
+  grad_clip: 0
+  instruments:
+    - soprano
+    - alto
+    - tenor
+    - bass
+  lr: 4.0e-04
+  patience: 6
+  reduce_factor: 0.95
+  target_instrument: null
+  num_epochs: 1000
+  num_steps: 1000
+  q: 0.95
+  coarse_loss_clip: true
+  ema_momentum: 0.999
+  optimizer: adamw
+  normalize: false # perform normalization on input of model (use the same for inference!)
+  other_fix: false # it's needed for checking on multisong dataset if other is actually instrumental
+  use_amp: true # enable or disable usage of mixed precision (float16) - usually it must be true
+
+augmentations:
+  enable: false # enable or disable all augmentations (to fast disable if needed)
+  loudness: true # randomly change loudness of each stem on the range (loudness_min; loudness_max)
+  loudness_min: 0.5
+  loudness_max: 1.5
+  mixup: false # mix several stems of same type with some probability (only works for dataset types: 1, 2, 3)
+  mixup_probs:
+    !!python/tuple # 2 additional stems of the same type (1st with prob 0.2, 2nd with prob 0.02)
+    - 0.2
+    - 0.02
+  mixup_loudness_min: 0.5
+  mixup_loudness_max: 1.5
+
+  # apply mp3 compression to mixture only (emulate downloading mp3 from internet)
+  mp3_compression_on_mixture: 0.01
+  mp3_compression_on_mixture_bitrate_min: 32
+  mp3_compression_on_mixture_bitrate_max: 320
+  mp3_compression_on_mixture_backend: "lameenc"
+
+  all:
+    channel_shuffle: 0.5 # Set 0 or lower to disable
+    random_inverse: 0.1 # inverse track (better lower probability)
+    random_polarity: 0.5 # polarity change (multiply waveform to -1)
+
+    mp3_compression: 0.01
+    mp3_compression_min_bitrate: 32
+    mp3_compression_max_bitrate: 320
+    mp3_compression_backend: "lameenc"
+
+    # pedalboard reverb block
+    pedalboard_reverb: 0.01
+    pedalboard_reverb_room_size_min: 0.1
+    pedalboard_reverb_room_size_max: 0.9
+    pedalboard_reverb_damping_min: 0.1
+    pedalboard_reverb_damping_max: 0.9
+    pedalboard_reverb_wet_level_min: 0.1
+    pedalboard_reverb_wet_level_max: 0.9
+    pedalboard_reverb_dry_level_min: 0.1
+    pedalboard_reverb_dry_level_max: 0.9
+    pedalboard_reverb_width_min: 0.9
+    pedalboard_reverb_width_max: 1.0
+
+    # pedalboard resample block
+    pedalboard_resample: 0.01
+    pedalboard_resample_target_sample_rate_min: 4000
+    pedalboard_resample_target_sample_rate_max: 44100
+    
+    pitch_shift: 0.1
+    pitch_shift_min_semitones: -1
+    pitch_shift_max_semitones: 1
+    
+    seven_band_parametric_eq: 0.1
+    seven_band_parametric_eq_min_gain_db: -9
+    seven_band_parametric_eq_max_gain_db: 9
+    
+
+inference:
+  batch_size: 8
+  dim_t: 256
+  num_overlap: 4
+  normalize: false