.

MDX23C_models/MDX23C-8KFFT-InstVoc_HQ.ckpt

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:49d51472769e34a2501cd1da782346a3212555c3a5619fc2c53507445528d816
-size 448101203

MDX23C_models/MDX23C-8KFFT-InstVoc_HQ_2.ckpt

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:7d960d8e40a458120412c1bd807e013d2dbca7b959cc9da2bbcb0eb203d1daea
-size 448093770

MDX23C_models/MDX23C-De-Reverb-aufr33-jarredou.ckpt

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:eae2471b707758d74db38ac1b1d5800e12f57c4e9d1ebbb2faf004b8e086e914
-size 448098867

MDX23C_models/MDX23C-DrumSep-aufr33-jarredou.ckpt

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:d2a4aa53eb584d21eead358a4e66d1882ad182911be018f052b5da73be9096d0
-size 437652699

MDX23C_models/MDX23C_D1581.ckpt

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:d7d4d2137f12764950205b095da20032fef1d41f077bacc8582f20ed40e8cb28
-size 183379219

MDX23C_models/config_dereverb_mdx23c.yaml

@@ -1,135 +0,0 @@
-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: 2
-  gradient_accumulation_steps: 1
-  grad_clip: 0
-  instruments:
-  - dry
-  - No dry
-  lr: 1.0e-06
-  patience: 4
-  reduce_factor: 0.93
-  target_instrument: null
-  num_epochs: 40
-  num_steps: 1000
-  q: 0.95
-  coarse_loss_clip: true
-  ema_momentum: 0.999
-  optimizer: adamw
-  read_metadata_procs: 8 # Number of processes to use during metadata reading for dataset. Can speed up metadata generation
-  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: 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)
-    - 0.2
-    - 0.02
-  mixup_loudness_min: 0.5
-  mixup_loudness_max: 1.5
-
-  all:
-    channel_shuffle: 0.5 # Set 0 or lower to disable
-    random_inverse: 0.05 # inverse track (better lower probability)
-    random_polarity: 0.5 # polarity change (multiply waveform to -1)
-
-    # pedalboard chorus block
-    pedalboard_chorus: 0.001
-    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: 0.01
-    pedalboard_chorus_mix_min: 0.1
-    pedalboard_chorus_mix_max: 0.9
-
-    # pedalboard phazer block
-    pedalboard_phazer: 0.001
-    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 pitch shift block
-    pedalboard_pitch_shift: 0.01
-    pedalboard_pitch_shift_semitones_min: -7
-    pedalboard_pitch_shift_semitones_max: 7
-
-    # pedalboard resample block
-    pedalboard_resample: 0.001
-    pedalboard_resample_target_sample_rate_min: 4000
-    pedalboard_resample_target_sample_rate_max: 44100
-
-    mp3_compression_min_bitrate: 32
-    mp3_compression_max_bitrate: 320
-    mp3_compression_backend: "lameenc"
-
-  dry:
-    # pedalboard distortion block
-    pedalboard_distortion: 0.001
-    pedalboard_distortion_drive_db_min: 1.0
-    pedalboard_distortion_drive_db_max: 25.0
-    
-    tanh_distortion: 0.05
-    tanh_distortion_min: 0.1
-    tanh_distortion_max: 0.7
-    # pedalboard bitcrash block
-    pedalboard_bitcrash: 0.005
-    pedalboard_bitcrash_bit_depth_min: 4
-    pedalboard_bitcrash_bit_depth_max: 16
-
-    seven_band_parametric_eq: 0.24
-    seven_band_parametric_eq_min_gain_db: -9
-    seven_band_parametric_eq_max_gain_db: 9
-    
-    gaussian_noise: 0.005
-    gaussian_noise_min_amplitude: 0.001
-    gaussian_noise_max_amplitude: 0.01
-    
-    time_stretch: 0.01
-    time_stretch_min_rate: 0.8
-    time_stretch_max_rate: 1.25
-  other:
-    seven_band_parametric_eq: 0.24
-    seven_band_parametric_eq_min_gain_db: -9
-    seven_band_parametric_eq_max_gain_db: 9
-
-inference:
-  batch_size: 2
-  dim_t: 256
-  num_overlap: 4

MDX23C_models/config_drumsep_mdx23c.yaml

@@ -1,87 +0,0 @@
-audio:
-  chunk_size: 130560
-  dim_f: 1024
-  dim_t: 256
-  hop_length: 512
-  n_fft: 2048
-  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: 12
-  gradient_accumulation_steps: 1
-  grad_clip: 0
-  instruments:
-  - kick
-  - snare
-  - toms
-  - hh
-  - ride
-  - crash
-  lr: 9.0e-05
-  patience: 30
-  reduce_factor: 0.95
-  target_instrument: null
-  num_epochs: 1000
-  num_steps: 1268
-  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
-  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)
-    - 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: 0.0
-  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: 0.01 # inverse track (better lower probability)
-    random_polarity: 0.5 # polarity change (multiply waveform to -1)
-    mp3_compression: 0.0
-    mp3_compression_min_bitrate: 32
-    mp3_compression_max_bitrate: 320
-    mp3_compression_backend: "lameenc"
-    pitch_shift: 0.1
-    pitch_shift_min_semitones: -3
-    pitch_shift_max_semitones: 3
-    seven_band_parametric_eq: 0.5
-    seven_band_parametric_eq_min_gain_db: -6
-    seven_band_parametric_eq_max_gain_db: 6
-    tanh_distortion: 0.2
-    tanh_distortion_min: 0.1
-    tanh_distortion_max: 0.5
-    
-inference:
-  batch_size: 1
-  dim_t: 256
-  num_overlap: 4

MDX23C_models/config_mdx23c_similarity.yaml

@@ -1,47 +0,0 @@
-audio:
-  chunk_size: 130560
-  dim_f: 1024
-  dim_t: 256
-  hop_length: 512
-  n_fft: 2048
-  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: 2
-  gradient_accumulation_steps: 3
-  grad_clip: 0
-  instruments:
-  - Similarity
-  - Difference
-  lr: 1.0
-  patience: 15
-  reduce_factor: 0.95
-  target_instrument: Similarity
-  num_epochs: 1000
-  num_steps: 2235
-  q: 0.95
-  coarse_loss_clip: true
-  ema_momentum: 0.999
-  optimizer: prodigy
-  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
-
-inference:
-  batch_size: 8
-  dim_t: 256
-  num_overlap: 8

MDX23C_models/model_2_stem_061321.yaml

@@ -1,36 +0,0 @@
-audio:
-  chunk_size: 260096
-  dim_f: 4096
-  dim_t: 256
-  hop_length: 2048
-  n_fft: 12288
-  num_channels: 2
-  sample_rate: 44100
-  min_mean_abs: 0.001
-model:
-  act: gelu
-  bottleneck_factor: 4
-  growth: 64
-  norm: InstanceNorm
-  num_blocks_per_scale: 2
-  num_channels: 128
-  num_scales: 5
-  num_subbands: 4
-  scale:
-  - 2
-  - 2
-  name: epoch_10.ckpt
-training:
-  batch_size: 16
-  grad_clip: 0
-  instruments:
-  - Vocals
-  - Instrumental
-  lr: 5.0e-05
-  target_instrument: null
-  num_epochs: 100
-  num_steps: 1000
-inference:
-  batch_size: 1
-  dim_t: 256
-  num_overlap: 8

MDX23C_models/model_2_stem_full_band_8k.yaml

@@ -1,43 +0,0 @@
-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
-  grad_clip: 0
-  instruments:
-  - Vocals
-  - Instrumental
-  lr: 1.0e-05
-  patience: 2
-  reduce_factor: 0.95
-  target_instrument: null
-  num_epochs: 1000
-  num_steps: 1000
-  augmentation: 1
-  augmentation_type: simple1
-  augmentation_mix: true
-  q: 0.95
-  coarse_loss_clip: true
-  ema_momentum: 0.999
-inference:
-  batch_size: 1
-  dim_t: 256
-  num_overlap: 8

MDX23C_models/model_mdx23c_ep_271_l1_freq_72.2383.ckpt

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:1227f8d97c7436004d03e46091427393abefebcc08ce53ef30082742c4e482f7
-size 437613512