deepgeek/CosyVoice
2
0
Fork 0
mirror of https://github.com/FunAudioLLM/CosyVoice.git synced 2026-02-04 17:39:25 +08:00
Code Issues Packages Projects Releases Wiki Activity

Merge pull request #323 from FunAudioLLM/hengwu.zty/zh_example

Browse Source 
change dropout_rate to float type
This commit is contained in:
Xiang Lyu
2024-08-28 19:00:45 +08:00
committed by GitHub
parent bb690d9d1e 567a136c0f
commit dcc943db43
2 changed files with 6 additions and 6 deletions
Download Patch File Download Diff File Expand all files Collapse all files

6
examples/magicdata-read/cosyvoice/conf/cosyvoice.fromscratch.yaml
View File

@@ -31,7 +31,7 @@ llm: !new:cosyvoice.llm.llm.TransformerLM
num_blocks: 3
dropout_rate: 0.1
positional_dropout_rate: 0.1
attention_dropout_rate: 0
attention_dropout_rate: 0.0
normalize_before: True
input_layer: 'linear'
pos_enc_layer_type: 'rel_pos_espnet'
@@ -49,7 +49,7 @@ llm: !new:cosyvoice.llm.llm.TransformerLM
num_blocks: 7
dropout_rate: 0.1
positional_dropout_rate: 0.1
attention_dropout_rate: 0
attention_dropout_rate: 0.0
input_layer: 'linear_legacy'
pos_enc_layer_type: 'rel_pos_espnet'
selfattention_layer_type: 'rel_selfattn'
@@ -97,7 +97,7 @@ flow: !new:cosyvoice.flow.flow.MaskedDiffWithXvec
in_channels: 320
out_channels: 80
channels: [256, 256]
dropout: 0
dropout: 0.0
attention_head_dim: 64
n_blocks: 4
num_mid_blocks: 8

6
examples/magicdata-read/cosyvoice/conf/cosyvoice.yaml
View File

@@ -31,7 +31,7 @@ llm: !new:cosyvoice.llm.llm.TransformerLM
num_blocks: 6
dropout_rate: 0.1
positional_dropout_rate: 0.1
attention_dropout_rate: 0
attention_dropout_rate: 0.0
normalize_before: True
input_layer: 'linear'
pos_enc_layer_type: 'rel_pos_espnet'
@@ -49,7 +49,7 @@ llm: !new:cosyvoice.llm.llm.TransformerLM
num_blocks: 14
dropout_rate: 0.1
positional_dropout_rate: 0.1
attention_dropout_rate: 0
attention_dropout_rate: 0.0
input_layer: 'linear_legacy'
pos_enc_layer_type: 'rel_pos_espnet'
selfattention_layer_type: 'rel_selfattn'
@@ -97,7 +97,7 @@ flow: !new:cosyvoice.flow.flow.MaskedDiffWithXvec
in_channels: 320
out_channels: 80
channels: [256, 256]
dropout: 0
dropout: 0.0
attention_head_dim: 64
n_blocks: 4
num_mid_blocks: 12