speechbrain
English
Vocoder
HiFIGAN
speech-synthesis
chaanks commited on
Commit
504e599
1 Parent(s): 53f22eb

Upload 2 files

Browse files
Files changed (2) hide show
  1. generator.ckpt +3 -0
  2. hyperparams.yaml +38 -0
generator.ckpt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:0c74ccc5e6a5c1162367a80d51300b40d2504ff33bd46f0409f59f61ff739d20
3
+ size 91090682
hyperparams.yaml ADDED
@@ -0,0 +1,38 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ vocab_size: 6001
2
+ embedding_dim: 1024
3
+
4
+ in_channels: 1024
5
+ out_channels: 1
6
+ resblock_type: "1"
7
+ resblock_dilation_sizes: [[1, 3, 5], [1, 3, 5], [1, 3, 5]]
8
+ resblock_kernel_sizes: [3, 7, 11]
9
+ upsample_kernel_sizes: [11, 8, 8, 4, 4]
10
+ upsample_initial_channel: 512
11
+ upsample_factors: [5, 4, 4, 2, 2]
12
+ inference_padding: 5
13
+ cond_channels: 0
14
+ conv_post_bias: True
15
+
16
+ generator: !new:speechbrain.lobes.models.HifiGAN.UnitHifiganGenerator
17
+ in_channels: !ref <in_channels>
18
+ out_channels: !ref <out_channels>
19
+ resblock_type: !ref <resblock_type>
20
+ resblock_dilation_sizes: !ref <resblock_dilation_sizes>
21
+ resblock_kernel_sizes: !ref <resblock_kernel_sizes>
22
+ upsample_kernel_sizes: !ref <upsample_kernel_sizes>
23
+ upsample_initial_channel: !ref <upsample_initial_channel>
24
+ upsample_factors: !ref <upsample_factors>
25
+ inference_padding: !ref <inference_padding>
26
+ cond_channels: !ref <cond_channels>
27
+ conv_post_bias: !ref <conv_post_bias>
28
+ vocab_size: !ref <vocab_size>
29
+ embedding_dim: !ref <embedding_dim>
30
+ duration_predictor: False
31
+ multi_speaker: False
32
+
33
+ modules:
34
+ generator: !ref <generator>
35
+
36
+ pretrainer: !new:speechbrain.utils.parameter_transfer.Pretrainer
37
+ loadables:
38
+ generator: !ref <generator>