Upload config.yaml with huggingface_hub

config.yaml

@@ -0,0 +1,208 @@
+audio:
+  chunk_size: 661500
+  dim_f: 1024
+  dim_t: 1101
+  hop_length: 441
+  n_fft: 2048
+  num_channels: 2
+  sample_rate: 44100
+  min_mean_abs: 0.001
+
+model:
+  dim: 512
+  depth: 8
+  stereo: true
+  num_stems: 1
+  time_transformer_depth: 1
+  freq_transformer_depth: 1
+  num_bands: 60
+  dim_head: 64
+  heads: 12
+  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: 3
+  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: 1
+  gradient_accumulation_steps: 2
+  grad_clip: 0
+  instruments:
+  - vocals
+  - instruments
+  lr: 1.0
+  patience: 2
+  reduce_factor: 0.9999999
+  target_instrument: vocals
+  num_epochs: 100000
+  num_steps: 10
+  augmentation: false # enable augmentations by audiomentations and pedalboard
+  augmentation_type: true
+  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: 0.25
+  augmentation_loudness_max: 1
+  q: 0.99
+  coarse_loss_clip: false
+  ema_momentum: 0.9995
+  optimizer: prodigy
+  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
+
+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: 0.01
+  loudness_max: 1.5
+  mixup: true # 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)
+    - 1
+    - 0.1
+  mixup_loudness_min: 0.25
+  mixup_loudness_max: 1
+
+  # apply mp3 compression to mixture only (emulate downloading mp3 from internet)
+  mp3_compression_on_mixture: 0
+  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: 0.7
+    pedalboard_reverb_room_size_min: 0.4
+    pedalboard_reverb_room_size_max: 0.9
+    pedalboard_reverb_damping_min: 0.3
+    pedalboard_reverb_damping_max: 0.9
+    pedalboard_reverb_wet_level_min: 0.4
+    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 chorus block
+    pedalboard_chorus: 0.3
+    pedalboard_chorus_rate_hz_min: 1.0
+    pedalboard_chorus_rate_hz_max: 7.0
+    pedalboard_chorus_depth_min: 0.25
+    pedalboard_chorus_depth_max: 0.95
+    pedalboard_chorus_centre_delay_ms_min: 3
+    pedalboard_chorus_centre_delay_ms_max: 10
+    pedalboard_chorus_feedback_min: 0.0
+    pedalboard_chorus_feedback_max: 0.5
+    pedalboard_chorus_mix_min: 0.1
+    pedalboard_chorus_mix_max: 0.9
+
+    # pedalboard phazer block
+    pedalboard_phazer: 0.2
+    pedalboard_phazer_rate_hz_min: 1.0
+    pedalboard_phazer_rate_hz_max: 10.0
+    pedalboard_phazer_depth_min: 0.25
+    pedalboard_phazer_depth_max: 0.95
+    pedalboard_phazer_centre_frequency_hz_min: 200
+    pedalboard_phazer_centre_frequency_hz_max: 12000
+    pedalboard_phazer_feedback_min: 0.0
+    pedalboard_phazer_feedback_max: 0.5
+    pedalboard_phazer_mix_min: 0.1
+    pedalboard_phazer_mix_max: 0.9
+
+    # pedalboard distortion block
+    pedalboard_distortion: 0.2
+    pedalboard_distortion_drive_db_min: 1.0
+    pedalboard_distortion_drive_db_max: 12 
+
+    # pedalboard pitch shift block
+    pedalboard_pitch_shift: 0
+    pedalboard_pitch_shift_semitones_min: -7
+    pedalboard_pitch_shift_semitones_max: 7
+
+    # pedalboard resample block
+    pedalboard_resample: 0.3
+    pedalboard_resample_target_sample_rate_min: 8000
+    pedalboard_resample_target_sample_rate_max: 44100
+
+    # pedalboard bitcrash block
+    pedalboard_bitcrash: 0.1
+    pedalboard_bitcrash_bit_depth_min: 8
+    pedalboard_bitcrash_bit_depth_max: 16
+
+    # pedalboard mp3 compressor block
+    pedalboard_mp3_compressor: 0
+    pedalboard_mp3_compressor_pedalboard_mp3_compressor_min: 0
+    pedalboard_mp3_compressor_pedalboard_mp3_compressor_max: 9.999
+
+  vocals:
+    pitch_shift: 0.7
+    pitch_shift_min_semitones: -6
+    pitch_shift_max_semitones: 6
+    seven_band_parametric_eq: 0.45
+    seven_band_parametric_eq_min_gain_db: -80
+    seven_band_parametric_eq_max_gain_db: 3
+    tanh_distortion: 0.3
+    tanh_distortion_min: 0.01
+    tanh_distortion_max: 0.7
+  other:
+    pitch_shift: 0.5
+    pitch_shift_min_semitones: -12
+    pitch_shift_max_semitones: 12
+    gaussian_noise: 0.5
+    gaussian_noise_min_amplitude: 0.01
+    gaussian_noise_max_amplitude: 0.5
+    time_stretch: 0.2
+    time_stretch_min_rate: 0.125
+    time_stretch_max_rate: 1.0
+
+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: 3.0
+  w_log_mag: 3.0
+  w_lin_mag: 2.0
+  w_phs: 3.0
+  mag_distance: "L1"
+
+inference:
+  batch_size: 4
+  dim_t: 1101
+  num_overlap: 4