ShiromiyaGamer
commited on
Commit
•
aef0324
1
Parent(s):
8fe63a8
Upload 2 files
Browse files- BS-Roformer_LargeV1.ckpt +3 -0
- config_bsrofoL.yaml +180 -0
BS-Roformer_LargeV1.ckpt
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:3d1334565bb9e4c08f4b3891d5ac39e269b511baf66024a3fd8a10112d52b945
|
3 |
+
size 740263433
|
config_bsrofoL.yaml
ADDED
@@ -0,0 +1,180 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
audio:
|
2 |
+
chunk_size: 352800
|
3 |
+
dim_f: 1024
|
4 |
+
dim_t: 801
|
5 |
+
hop_length: 441
|
6 |
+
n_fft: 2048
|
7 |
+
num_channels: 2
|
8 |
+
sample_rate: 44100
|
9 |
+
min_mean_abs: 0.000
|
10 |
+
|
11 |
+
model:
|
12 |
+
dim: 512
|
13 |
+
depth: 16
|
14 |
+
stereo: true
|
15 |
+
num_stems: 1
|
16 |
+
time_transformer_depth: 1
|
17 |
+
freq_transformer_depth: 1
|
18 |
+
linear_transformer_depth: 0
|
19 |
+
freqs_per_bands: !!python/tuple
|
20 |
+
- 2
|
21 |
+
- 2
|
22 |
+
- 2
|
23 |
+
- 2
|
24 |
+
- 2
|
25 |
+
- 2
|
26 |
+
- 2
|
27 |
+
- 2
|
28 |
+
- 2
|
29 |
+
- 2
|
30 |
+
- 2
|
31 |
+
- 2
|
32 |
+
- 2
|
33 |
+
- 2
|
34 |
+
- 2
|
35 |
+
- 2
|
36 |
+
- 2
|
37 |
+
- 2
|
38 |
+
- 2
|
39 |
+
- 2
|
40 |
+
- 2
|
41 |
+
- 2
|
42 |
+
- 2
|
43 |
+
- 2
|
44 |
+
- 4
|
45 |
+
- 4
|
46 |
+
- 4
|
47 |
+
- 4
|
48 |
+
- 4
|
49 |
+
- 4
|
50 |
+
- 4
|
51 |
+
- 4
|
52 |
+
- 4
|
53 |
+
- 4
|
54 |
+
- 4
|
55 |
+
- 4
|
56 |
+
- 12
|
57 |
+
- 12
|
58 |
+
- 12
|
59 |
+
- 12
|
60 |
+
- 12
|
61 |
+
- 12
|
62 |
+
- 12
|
63 |
+
- 12
|
64 |
+
- 24
|
65 |
+
- 24
|
66 |
+
- 24
|
67 |
+
- 24
|
68 |
+
- 24
|
69 |
+
- 24
|
70 |
+
- 24
|
71 |
+
- 24
|
72 |
+
- 48
|
73 |
+
- 48
|
74 |
+
- 48
|
75 |
+
- 48
|
76 |
+
- 48
|
77 |
+
- 48
|
78 |
+
- 48
|
79 |
+
- 48
|
80 |
+
- 128
|
81 |
+
- 129
|
82 |
+
dim_head: 64
|
83 |
+
heads: 8
|
84 |
+
attn_dropout: 0.1
|
85 |
+
ff_dropout: 0.1
|
86 |
+
flash_attn: true
|
87 |
+
dim_freqs_in: 1025
|
88 |
+
stft_n_fft: 2048
|
89 |
+
stft_hop_length: 441
|
90 |
+
stft_win_length: 2048
|
91 |
+
stft_normalized: false
|
92 |
+
mask_estimator_depth: 2
|
93 |
+
multi_stft_resolution_loss_weight: 1.0
|
94 |
+
multi_stft_resolutions_window_sizes: !!python/tuple
|
95 |
+
- 4096
|
96 |
+
- 2048
|
97 |
+
- 1024
|
98 |
+
- 512
|
99 |
+
- 256
|
100 |
+
multi_stft_hop_size: 147
|
101 |
+
multi_stft_normalized: False
|
102 |
+
|
103 |
+
training:
|
104 |
+
batch_size: 1
|
105 |
+
gradient_accumulation_steps: 1
|
106 |
+
grad_clip: 0
|
107 |
+
instruments:
|
108 |
+
- vocals
|
109 |
+
- other
|
110 |
+
lr: 2.0e-07
|
111 |
+
patience: 2
|
112 |
+
reduce_factor: 0.95
|
113 |
+
target_instrument: vocals
|
114 |
+
num_epochs: 1000
|
115 |
+
num_steps: 5000
|
116 |
+
q: 0.95
|
117 |
+
coarse_loss_clip: true
|
118 |
+
ema_momentum: 0.999
|
119 |
+
optimizer: adamw8bit
|
120 |
+
other_fix: true # it's needed for checking on multisong dataset if other is actually instrumental
|
121 |
+
use_amp: true # enable or disable usage of mixed precision (float16) - usually it must be true
|
122 |
+
|
123 |
+
augmentations:
|
124 |
+
enable: true # enable or disable all augmentations (to fast disable if needed)
|
125 |
+
loudness: true # randomly change loudness of each stem on the range (loudness_min; loudness_max)
|
126 |
+
loudness_min: 0.5
|
127 |
+
loudness_max: 1.5
|
128 |
+
mixup: false # mix several stems of same type with some probability (only works for dataset types: 1, 2, 3)
|
129 |
+
mixup_probs:
|
130 |
+
!!python/tuple # 2 additional stems of the same type (1st with prob 0.2, 2nd with prob 0.02)
|
131 |
+
- 0.2
|
132 |
+
- 0.02
|
133 |
+
mixup_loudness_min: 0.25
|
134 |
+
mixup_loudness_max: 1.75
|
135 |
+
all:
|
136 |
+
channel_shuffle: 0.5 # Set 0 or lower to disable
|
137 |
+
random_inverse: 0.1 # inverse track (better lower probability)
|
138 |
+
random_polarity: 0.5 # polarity change (multiply waveform to -1)
|
139 |
+
mp3_compression: 0.05
|
140 |
+
mp3_compression_min_bitrate: 32
|
141 |
+
mp3_compression_max_bitrate: 320
|
142 |
+
mp3_compression_backend: "lameenc"
|
143 |
+
|
144 |
+
|
145 |
+
# pedalboard resample block
|
146 |
+
pedalboard_resample: 0.001
|
147 |
+
pedalboard_resample_target_sample_rate_min: 4000
|
148 |
+
pedalboard_resample_target_sample_rate_max: 44100
|
149 |
+
|
150 |
+
# pedalboard mp3 compressor block
|
151 |
+
pedalboard_mp3_compressor: 0.005
|
152 |
+
pedalboard_mp3_compressor_pedalboard_mp3_compressor_min: 0
|
153 |
+
pedalboard_mp3_compressor_pedalboard_mp3_compressor_max: 9.999
|
154 |
+
|
155 |
+
vocals:
|
156 |
+
pitch_shift: 0.25
|
157 |
+
pitch_shift_min_semitones: -7
|
158 |
+
pitch_shift_max_semitones: 7
|
159 |
+
seven_band_parametric_eq: 0.25
|
160 |
+
seven_band_parametric_eq_min_gain_db: -9
|
161 |
+
seven_band_parametric_eq_max_gain_db: 9
|
162 |
+
tanh_distortion: 0.01
|
163 |
+
tanh_distortion_min: 0.1
|
164 |
+
tanh_distortion_max: 0.7
|
165 |
+
pedalboard_reverb: 0.01
|
166 |
+
pedalboard_reverb_room_size_min: 0.1
|
167 |
+
pedalboard_reverb_room_size_max: 0.9
|
168 |
+
pedalboard_reverb_damping_min: 0.1
|
169 |
+
pedalboard_reverb_damping_max: 0.9
|
170 |
+
pedalboard_reverb_wet_level_min: 0.1
|
171 |
+
pedalboard_reverb_wet_level_max: 0.9
|
172 |
+
pedalboard_reverb_dry_level_min: 0.1
|
173 |
+
pedalboard_reverb_dry_level_max: 0.9
|
174 |
+
pedalboard_reverb_width_min: 0.9
|
175 |
+
pedalboard_reverb_width_max: 1.0
|
176 |
+
|
177 |
+
inference:
|
178 |
+
batch_size: 2
|
179 |
+
dim_t: 1101
|
180 |
+
num_overlap: 2
|