Upload config.yaml
Browse files- misc/New folder/config.yaml +16 -10
misc/New folder/config.yaml
CHANGED
|
@@ -1,3 +1,12 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
|
| 2 |
model:
|
| 3 |
dim: 384
|
|
@@ -13,7 +22,7 @@ model:
|
|
| 13 |
ff_dropout: 0
|
| 14 |
flash_attn: True
|
| 15 |
dim_freqs_in: 1025
|
| 16 |
-
sample_rate: 44100
|
| 17 |
stft_n_fft: 2048
|
| 18 |
stft_hop_length: 441
|
| 19 |
stft_win_length: 2048
|
|
@@ -30,16 +39,13 @@ model:
|
|
| 30 |
multi_stft_normalized: False
|
| 31 |
|
| 32 |
training:
|
| 33 |
-
batch_size: 1
|
| 34 |
-
gradient_accumulation_steps: 1
|
| 35 |
-
grad_clip: 0
|
| 36 |
instruments:
|
| 37 |
-
-
|
| 38 |
-
-
|
| 39 |
-
target_instrument:
|
| 40 |
-
use_amp: true
|
| 41 |
|
| 42 |
inference:
|
| 43 |
-
batch_size:
|
| 44 |
dim_t: 1101
|
| 45 |
-
num_overlap:
|
|
|
|
|
|
| 1 |
+
audio:
|
| 2 |
+
chunk_size: 352800
|
| 3 |
+
dim_f: 1024
|
| 4 |
+
dim_t: 256
|
| 5 |
+
hop_length: 441
|
| 6 |
+
n_fft: 2048
|
| 7 |
+
num_channels: 2
|
| 8 |
+
sample_rate: 44100
|
| 9 |
+
min_mean_abs: 0.001
|
| 10 |
|
| 11 |
model:
|
| 12 |
dim: 384
|
|
|
|
| 22 |
ff_dropout: 0
|
| 23 |
flash_attn: True
|
| 24 |
dim_freqs_in: 1025
|
| 25 |
+
sample_rate: 44100 # needed for mel filter bank from librosa
|
| 26 |
stft_n_fft: 2048
|
| 27 |
stft_hop_length: 441
|
| 28 |
stft_win_length: 2048
|
|
|
|
| 39 |
multi_stft_normalized: False
|
| 40 |
|
| 41 |
training:
|
|
|
|
|
|
|
|
|
|
| 42 |
instruments:
|
| 43 |
+
- Vocals
|
| 44 |
+
- Instrumental
|
| 45 |
+
target_instrument: Vocals
|
|
|
|
| 46 |
|
| 47 |
inference:
|
| 48 |
+
batch_size: 1
|
| 49 |
dim_t: 1101
|
| 50 |
+
num_overlap: 1
|
| 51 |
+
chunk_size: 352800
|