Skip to content
体验新版
项目
组织
正在加载...
登录
切换导航
打开侧边栏
PaddlePaddle
DeepSpeech
提交
1169ffa4
D
DeepSpeech
项目概览
PaddlePaddle
/
DeepSpeech
大约 1 年 前同步成功
通知
207
Star
8425
Fork
1598
代码
文件
提交
分支
Tags
贡献者
分支图
Diff
Issue
245
列表
看板
标记
里程碑
合并请求
3
Wiki
0
Wiki
分析
仓库
DevOps
项目成员
Pages
D
DeepSpeech
项目概览
项目概览
详情
发布
仓库
仓库
文件
提交
分支
标签
贡献者
分支图
比较
Issue
245
Issue
245
列表
看板
标记
里程碑
合并请求
3
合并请求
3
Pages
分析
分析
仓库分析
DevOps
Wiki
0
Wiki
成员
成员
收起侧边栏
关闭侧边栏
动态
分支图
创建新Issue
提交
Issue看板
体验新版 GitCode,发现更多精彩内容 >>
提交
1169ffa4
编写于
2月 17, 2022
作者:
J
Junkun
浏览文件
操作
浏览文件
下载
电子邮件补丁
差异文件
add config files
上级
70166c20
变更
8
显示空白变更内容
内联
并排
Showing
8 changed file
with
720 addition
and
0 deletion
+720
-0
examples/mustc/st1/conf/transformer_de.yaml
examples/mustc/st1/conf/transformer_de.yaml
+90
-0
examples/mustc/st1/conf/transformer_es.yaml
examples/mustc/st1/conf/transformer_es.yaml
+90
-0
examples/mustc/st1/conf/transformer_fr.yaml
examples/mustc/st1/conf/transformer_fr.yaml
+90
-0
examples/mustc/st1/conf/transformer_it.yaml
examples/mustc/st1/conf/transformer_it.yaml
+90
-0
examples/mustc/st1/conf/transformer_nl.yaml
examples/mustc/st1/conf/transformer_nl.yaml
+90
-0
examples/mustc/st1/conf/transformer_pt.yaml
examples/mustc/st1/conf/transformer_pt.yaml
+90
-0
examples/mustc/st1/conf/transformer_ro.yaml
examples/mustc/st1/conf/transformer_ro.yaml
+90
-0
examples/mustc/st1/conf/transformer_ru.yaml
examples/mustc/st1/conf/transformer_ru.yaml
+90
-0
未找到文件。
examples/mustc/st1/conf/transformer_de.yaml
0 → 100644
浏览文件 @
1169ffa4
# https://yaml.org/type/float.html
###########################################
# Data #
###########################################
train_manifest
:
data/manifest.de.train
dev_manifest
:
data/manifest.de.dev
test_manifest
:
data/manifest.de.test
###########################################
# Dataloader #
###########################################
vocab_filepath
:
data/lang_1spm/train_sp.en-de.de_bpe8000_units_tc.txt
unit_type
:
'
spm'
spm_model_prefix
:
data/lang_1spm/train_sp.en-de.de_bpe8000_tc
mean_std_filepath
:
"
"
# preprocess_config: conf/augmentation.json
batch_size
:
20
feat_dim
:
83
stride_ms
:
10.0
window_ms
:
25.0
sortagrad
:
0
# Feed samples from shortest to longest ; -1: enabled for all epochs, 0: disabled, other: enabled for 'other' epochs
maxlen_in
:
512
# if input length > maxlen-in, batchsize is automatically reduced
maxlen_out
:
150
# if output length > maxlen-out, batchsize is automatically reduced
minibatches
:
0
# for debug
batch_count
:
auto
batch_bins
:
0
batch_frames_in
:
0
batch_frames_out
:
0
batch_frames_inout
:
0
preprocess_config
:
num_workers
:
0
subsampling_factor
:
1
num_encs
:
1
############################################
# Network Architecture #
############################################
cmvn_file
:
None
cmvn_file_type
:
"
json"
# encoder related
encoder
:
transformer
encoder_conf
:
output_size
:
256
# dimension of attention
attention_heads
:
4
linear_units
:
2048
# the number of units of position-wise feed forward
num_blocks
:
12
# the number of encoder blocks
dropout_rate
:
0.1
positional_dropout_rate
:
0.1
attention_dropout_rate
:
0.0
input_layer
:
conv2d
# encoder input type, you can chose conv2d, conv2d6 and conv2d8
normalize_before
:
true
# decoder related
decoder
:
transformer
decoder_conf
:
attention_heads
:
4
linear_units
:
2048
num_blocks
:
6
dropout_rate
:
0.1
positional_dropout_rate
:
0.1
self_attention_dropout_rate
:
0.0
src_attention_dropout_rate
:
0.0
# hybrid CTC/attention
model_conf
:
asr_weight
:
0.0
ctc_weight
:
0.0
lsm_weight
:
0.1
# label smoothing option
length_normalized_loss
:
false
###########################################
# Training #
###########################################
n_epoch
:
40
accum_grad
:
2
global_grad_clip
:
5.0
optim
:
adam
optim_conf
:
lr
:
2.5
weight_decay
:
0.
scheduler
:
noam
scheduler_conf
:
warmup_steps
:
25000
lr_decay
:
1.0
log_interval
:
50
checkpoint
:
kbest_n
:
50
latest_n
:
5
\ No newline at end of file
examples/mustc/st1/conf/transformer_es.yaml
0 → 100644
浏览文件 @
1169ffa4
# https://yaml.org/type/float.html
###########################################
# Data #
###########################################
train_manifest
:
data/manifest.es.train
dev_manifest
:
data/manifest.es.dev
test_manifest
:
data/manifest.es.test
###########################################
# Dataloader #
###########################################
vocab_filepath
:
data/lang_1spm/train_sp.en-es.es_bpe8000_units_tc.txt
unit_type
:
'
spm'
spm_model_prefix
:
data/lang_1spm/train_sp.en-es.es_bpe8000_tc
mean_std_filepath
:
"
"
# preprocess_config: conf/augmentation.json
batch_size
:
20
feat_dim
:
83
stride_ms
:
10.0
window_ms
:
25.0
sortagrad
:
0
# Feed samples from shortest to longest ; -1: enabled for all epochs, 0: disabled, other: enabled for 'other' epochs
maxlen_in
:
512
# if input length > maxlen-in, batchsize is automatically reduced
maxlen_out
:
150
# if output length > maxlen-out, batchsize is automatically reduced
minibatches
:
0
# for debug
batch_count
:
auto
batch_bins
:
0
batch_frames_in
:
0
batch_frames_out
:
0
batch_frames_inout
:
0
preprocess_config
:
num_workers
:
0
subsampling_factor
:
1
num_encs
:
1
############################################
# Network Architecture #
############################################
cmvn_file
:
None
cmvn_file_type
:
"
json"
# encoder related
encoder
:
transformer
encoder_conf
:
output_size
:
256
# dimension of attention
attention_heads
:
4
linear_units
:
2048
# the number of units of position-wise feed forward
num_blocks
:
12
# the number of encoder blocks
dropout_rate
:
0.1
positional_dropout_rate
:
0.1
attention_dropout_rate
:
0.0
input_layer
:
conv2d
# encoder input type, you can chose conv2d, conv2d6 and conv2d8
normalize_before
:
true
# decoder related
decoder
:
transformer
decoder_conf
:
attention_heads
:
4
linear_units
:
2048
num_blocks
:
6
dropout_rate
:
0.1
positional_dropout_rate
:
0.1
self_attention_dropout_rate
:
0.0
src_attention_dropout_rate
:
0.0
# hybrid CTC/attention
model_conf
:
asr_weight
:
0.0
ctc_weight
:
0.0
lsm_weight
:
0.1
# label smoothing option
length_normalized_loss
:
false
###########################################
# Training #
###########################################
n_epoch
:
40
accum_grad
:
2
global_grad_clip
:
5.0
optim
:
adam
optim_conf
:
lr
:
2.5
weight_decay
:
0.
scheduler
:
noam
scheduler_conf
:
warmup_steps
:
25000
lr_decay
:
1.0
log_interval
:
50
checkpoint
:
kbest_n
:
50
latest_n
:
5
\ No newline at end of file
examples/mustc/st1/conf/transformer_fr.yaml
0 → 100644
浏览文件 @
1169ffa4
# https://yaml.org/type/float.html
###########################################
# Data #
###########################################
train_manifest
:
data/manifest.fr.train
dev_manifest
:
data/manifest.fr.dev
test_manifest
:
data/manifest.fr.test
###########################################
# Dataloader #
###########################################
vocab_filepath
:
data/lang_1spm/train_sp.en-fr.fr_bpe8000_units_tc.txt
unit_type
:
'
spm'
spm_model_prefix
:
data/lang_1spm/train_sp.en-fr.fr_bpe8000_tc
mean_std_filepath
:
"
"
# preprocess_config: conf/augmentation.json
batch_size
:
20
feat_dim
:
83
stride_ms
:
10.0
window_ms
:
25.0
sortagrad
:
0
# Feed samples from shortest to longest ; -1: enabled for all epochs, 0: disabled, other: enabled for 'other' epochs
maxlen_in
:
512
# if input length > maxlen-in, batchsize is automatically reduced
maxlen_out
:
150
# if output length > maxlen-out, batchsize is automatically reduced
minibatches
:
0
# for debug
batch_count
:
auto
batch_bins
:
0
batch_frames_in
:
0
batch_frames_out
:
0
batch_frames_inout
:
0
preprocess_config
:
num_workers
:
0
subsampling_factor
:
1
num_encs
:
1
############################################
# Network Architecture #
############################################
cmvn_file
:
None
cmvn_file_type
:
"
json"
# encoder related
encoder
:
transformer
encoder_conf
:
output_size
:
256
# dimension of attention
attention_heads
:
4
linear_units
:
2048
# the number of units of position-wise feed forward
num_blocks
:
12
# the number of encoder blocks
dropout_rate
:
0.1
positional_dropout_rate
:
0.1
attention_dropout_rate
:
0.0
input_layer
:
conv2d
# encoder input type, you can chose conv2d, conv2d6 and conv2d8
normalize_before
:
true
# decoder related
decoder
:
transformer
decoder_conf
:
attention_heads
:
4
linear_units
:
2048
num_blocks
:
6
dropout_rate
:
0.1
positional_dropout_rate
:
0.1
self_attention_dropout_rate
:
0.0
src_attention_dropout_rate
:
0.0
# hybrid CTC/attention
model_conf
:
asr_weight
:
0.0
ctc_weight
:
0.0
lsm_weight
:
0.1
# label smoothing option
length_normalized_loss
:
false
###########################################
# Training #
###########################################
n_epoch
:
40
accum_grad
:
2
global_grad_clip
:
5.0
optim
:
adam
optim_conf
:
lr
:
2.5
weight_decay
:
0.
scheduler
:
noam
scheduler_conf
:
warmup_steps
:
25000
lr_decay
:
1.0
log_interval
:
50
checkpoint
:
kbest_n
:
50
latest_n
:
5
\ No newline at end of file
examples/mustc/st1/conf/transformer_it.yaml
0 → 100644
浏览文件 @
1169ffa4
# https://yaml.org/type/float.html
###########################################
# Data #
###########################################
train_manifest
:
data/manifest.it.train
dev_manifest
:
data/manifest.it.dev
test_manifest
:
data/manifest.it.test
###########################################
# Dataloader #
###########################################
vocab_filepath
:
data/lang_1spm/train_sp.en-it.it_bpe8000_units_tc.txt
unit_type
:
'
spm'
spm_model_prefix
:
data/lang_1spm/train_sp.en-it.it_bpe8000_tc
mean_std_filepath
:
"
"
# preprocess_config: conf/augmentation.json
batch_size
:
20
feat_dim
:
83
stride_ms
:
10.0
window_ms
:
25.0
sortagrad
:
0
# Feed samples from shortest to longest ; -1: enabled for all epochs, 0: disabled, other: enabled for 'other' epochs
maxlen_in
:
512
# if input length > maxlen-in, batchsize is automatically reduced
maxlen_out
:
150
# if output length > maxlen-out, batchsize is automatically reduced
minibatches
:
0
# for debug
batch_count
:
auto
batch_bins
:
0
batch_frames_in
:
0
batch_frames_out
:
0
batch_frames_inout
:
0
preprocess_config
:
num_workers
:
0
subsampling_factor
:
1
num_encs
:
1
############################################
# Network Architecture #
############################################
cmvn_file
:
None
cmvn_file_type
:
"
json"
# encoder related
encoder
:
transformer
encoder_conf
:
output_size
:
256
# dimension of attention
attention_heads
:
4
linear_units
:
2048
# the number of units of position-wise feed forward
num_blocks
:
12
# the number of encoder blocks
dropout_rate
:
0.1
positional_dropout_rate
:
0.1
attention_dropout_rate
:
0.0
input_layer
:
conv2d
# encoder input type, you can chose conv2d, conv2d6 and conv2d8
normalize_before
:
true
# decoder related
decoder
:
transformer
decoder_conf
:
attention_heads
:
4
linear_units
:
2048
num_blocks
:
6
dropout_rate
:
0.1
positional_dropout_rate
:
0.1
self_attention_dropout_rate
:
0.0
src_attention_dropout_rate
:
0.0
# hybrid CTC/attention
model_conf
:
asr_weight
:
0.0
ctc_weight
:
0.0
lsm_weight
:
0.1
# label smoothing option
length_normalized_loss
:
false
###########################################
# Training #
###########################################
n_epoch
:
40
accum_grad
:
2
global_grad_clip
:
5.0
optim
:
adam
optim_conf
:
lr
:
2.5
weight_decay
:
0.
scheduler
:
noam
scheduler_conf
:
warmup_steps
:
25000
lr_decay
:
1.0
log_interval
:
50
checkpoint
:
kbest_n
:
50
latest_n
:
5
\ No newline at end of file
examples/mustc/st1/conf/transformer_nl.yaml
0 → 100644
浏览文件 @
1169ffa4
# https://yaml.org/type/float.html
###########################################
# Data #
###########################################
train_manifest
:
data/manifest.nl.train
dev_manifest
:
data/manifest.nl.dev
test_manifest
:
data/manifest.nl.test
###########################################
# Dataloader #
###########################################
vocab_filepath
:
data/lang_1spm/train_sp.en-nl.nl_bpe8000_units_tc.txt
unit_type
:
'
spm'
spm_model_prefix
:
data/lang_1spm/train_sp.en-nl.nl_bpe8000_tc
mean_std_filepath
:
"
"
# preprocess_config: conf/augmentation.json
batch_size
:
20
feat_dim
:
83
stride_ms
:
10.0
window_ms
:
25.0
sortagrad
:
0
# Feed samples from shortest to longest ; -1: enabled for all epochs, 0: disabled, other: enabled for 'other' epochs
maxlen_in
:
512
# if input length > maxlen-in, batchsize is automatically reduced
maxlen_out
:
150
# if output length > maxlen-out, batchsize is automatically reduced
minibatches
:
0
# for debug
batch_count
:
auto
batch_bins
:
0
batch_frames_in
:
0
batch_frames_out
:
0
batch_frames_inout
:
0
preprocess_config
:
num_workers
:
0
subsampling_factor
:
1
num_encs
:
1
############################################
# Network Architecture #
############################################
cmvn_file
:
None
cmvn_file_type
:
"
json"
# encoder related
encoder
:
transformer
encoder_conf
:
output_size
:
256
# dimension of attention
attention_heads
:
4
linear_units
:
2048
# the number of units of position-wise feed forward
num_blocks
:
12
# the number of encoder blocks
dropout_rate
:
0.1
positional_dropout_rate
:
0.1
attention_dropout_rate
:
0.0
input_layer
:
conv2d
# encoder input type, you can chose conv2d, conv2d6 and conv2d8
normalize_before
:
true
# decoder related
decoder
:
transformer
decoder_conf
:
attention_heads
:
4
linear_units
:
2048
num_blocks
:
6
dropout_rate
:
0.1
positional_dropout_rate
:
0.1
self_attention_dropout_rate
:
0.0
src_attention_dropout_rate
:
0.0
# hybrid CTC/attention
model_conf
:
asr_weight
:
0.0
ctc_weight
:
0.0
lsm_weight
:
0.1
# label smoothing option
length_normalized_loss
:
false
###########################################
# Training #
###########################################
n_epoch
:
40
accum_grad
:
2
global_grad_clip
:
5.0
optim
:
adam
optim_conf
:
lr
:
2.5
weight_decay
:
0.
scheduler
:
noam
scheduler_conf
:
warmup_steps
:
25000
lr_decay
:
1.0
log_interval
:
50
checkpoint
:
kbest_n
:
50
latest_n
:
5
\ No newline at end of file
examples/mustc/st1/conf/transformer_pt.yaml
0 → 100644
浏览文件 @
1169ffa4
# https://yaml.org/type/float.html
###########################################
# Data #
###########################################
train_manifest
:
data/manifest.pt.train
dev_manifest
:
data/manifest.pt.dev
test_manifest
:
data/manifest.pt.test
###########################################
# Dataloader #
###########################################
vocab_filepath
:
data/lang_1spm/train_sp.en-pt.pt_bpe8000_units_tc.txt
unit_type
:
'
spm'
spm_model_prefix
:
data/lang_1spm/train_sp.en-pt.pt_bpe8000_tc
mean_std_filepath
:
"
"
# preprocess_config: conf/augmentation.json
batch_size
:
20
feat_dim
:
83
stride_ms
:
10.0
window_ms
:
25.0
sortagrad
:
0
# Feed samples from shortest to longest ; -1: enabled for all epochs, 0: disabled, other: enabled for 'other' epochs
maxlen_in
:
512
# if input length > maxlen-in, batchsize is automatically reduced
maxlen_out
:
150
# if output length > maxlen-out, batchsize is automatically reduced
minibatches
:
0
# for debug
batch_count
:
auto
batch_bins
:
0
batch_frames_in
:
0
batch_frames_out
:
0
batch_frames_inout
:
0
preprocess_config
:
num_workers
:
0
subsampling_factor
:
1
num_encs
:
1
############################################
# Network Architecture #
############################################
cmvn_file
:
None
cmvn_file_type
:
"
json"
# encoder related
encoder
:
transformer
encoder_conf
:
output_size
:
256
# dimension of attention
attention_heads
:
4
linear_units
:
2048
# the number of units of position-wise feed forward
num_blocks
:
12
# the number of encoder blocks
dropout_rate
:
0.1
positional_dropout_rate
:
0.1
attention_dropout_rate
:
0.0
input_layer
:
conv2d
# encoder input type, you can chose conv2d, conv2d6 and conv2d8
normalize_before
:
true
# decoder related
decoder
:
transformer
decoder_conf
:
attention_heads
:
4
linear_units
:
2048
num_blocks
:
6
dropout_rate
:
0.1
positional_dropout_rate
:
0.1
self_attention_dropout_rate
:
0.0
src_attention_dropout_rate
:
0.0
# hybrid CTC/attention
model_conf
:
asr_weight
:
0.0
ctc_weight
:
0.0
lsm_weight
:
0.1
# label smoothing option
length_normalized_loss
:
false
###########################################
# Training #
###########################################
n_epoch
:
40
accum_grad
:
2
global_grad_clip
:
5.0
optim
:
adam
optim_conf
:
lr
:
2.5
weight_decay
:
0.
scheduler
:
noam
scheduler_conf
:
warmup_steps
:
25000
lr_decay
:
1.0
log_interval
:
50
checkpoint
:
kbest_n
:
50
latest_n
:
5
\ No newline at end of file
examples/mustc/st1/conf/transformer_ro.yaml
0 → 100644
浏览文件 @
1169ffa4
# https://yaml.org/type/float.html
###########################################
# Data #
###########################################
train_manifest
:
data/manifest.ro.train
dev_manifest
:
data/manifest.ro.dev
test_manifest
:
data/manifest.ro.test
###########################################
# Dataloader #
###########################################
vocab_filepath
:
data/lang_1spm/train_sp.en-ro.ro_bpe8000_units_tc.txt
unit_type
:
'
spm'
spm_model_prefix
:
data/lang_1spm/train_sp.en-ro.ro_bpe8000_tc
mean_std_filepath
:
"
"
# preprocess_config: conf/augmentation.json
batch_size
:
20
feat_dim
:
83
stride_ms
:
10.0
window_ms
:
25.0
sortagrad
:
0
# Feed samples from shortest to longest ; -1: enabled for all epochs, 0: disabled, other: enabled for 'other' epochs
maxlen_in
:
512
# if input length > maxlen-in, batchsize is automatically reduced
maxlen_out
:
150
# if output length > maxlen-out, batchsize is automatically reduced
minibatches
:
0
# for debug
batch_count
:
auto
batch_bins
:
0
batch_frames_in
:
0
batch_frames_out
:
0
batch_frames_inout
:
0
preprocess_config
:
num_workers
:
0
subsampling_factor
:
1
num_encs
:
1
############################################
# Network Architecture #
############################################
cmvn_file
:
None
cmvn_file_type
:
"
json"
# encoder related
encoder
:
transformer
encoder_conf
:
output_size
:
256
# dimension of attention
attention_heads
:
4
linear_units
:
2048
# the number of units of position-wise feed forward
num_blocks
:
12
# the number of encoder blocks
dropout_rate
:
0.1
positional_dropout_rate
:
0.1
attention_dropout_rate
:
0.0
input_layer
:
conv2d
# encoder input type, you can chose conv2d, conv2d6 and conv2d8
normalize_before
:
true
# decoder related
decoder
:
transformer
decoder_conf
:
attention_heads
:
4
linear_units
:
2048
num_blocks
:
6
dropout_rate
:
0.1
positional_dropout_rate
:
0.1
self_attention_dropout_rate
:
0.0
src_attention_dropout_rate
:
0.0
# hybrid CTC/attention
model_conf
:
asr_weight
:
0.0
ctc_weight
:
0.0
lsm_weight
:
0.1
# label smoothing option
length_normalized_loss
:
false
###########################################
# Training #
###########################################
n_epoch
:
40
accum_grad
:
2
global_grad_clip
:
5.0
optim
:
adam
optim_conf
:
lr
:
2.5
weight_decay
:
0.
scheduler
:
noam
scheduler_conf
:
warmup_steps
:
25000
lr_decay
:
1.0
log_interval
:
50
checkpoint
:
kbest_n
:
50
latest_n
:
5
\ No newline at end of file
examples/mustc/st1/conf/transformer_ru.yaml
0 → 100644
浏览文件 @
1169ffa4
# https://yaml.org/type/float.html
###########################################
# Data #
###########################################
train_manifest
:
data/manifest.ru.train
dev_manifest
:
data/manifest.ru.dev
test_manifest
:
data/manifest.ru.test
###########################################
# Dataloader #
###########################################
vocab_filepath
:
data/lang_1spm/train_sp.en-ru.ru_bpe8000_units_tc.txt
unit_type
:
'
spm'
spm_model_prefix
:
data/lang_1spm/train_sp.en-ru.ru_bpe8000_tc
mean_std_filepath
:
"
"
# preprocess_config: conf/augmentation.json
batch_size
:
20
feat_dim
:
83
stride_ms
:
10.0
window_ms
:
25.0
sortagrad
:
0
# Feed samples from shortest to longest ; -1: enabled for all epochs, 0: disabled, other: enabled for 'other' epochs
maxlen_in
:
512
# if input length > maxlen-in, batchsize is automatically reduced
maxlen_out
:
150
# if output length > maxlen-out, batchsize is automatically reduced
minibatches
:
0
# for debug
batch_count
:
auto
batch_bins
:
0
batch_frames_in
:
0
batch_frames_out
:
0
batch_frames_inout
:
0
preprocess_config
:
num_workers
:
0
subsampling_factor
:
1
num_encs
:
1
############################################
# Network Architecture #
############################################
cmvn_file
:
None
cmvn_file_type
:
"
json"
# encoder related
encoder
:
transformer
encoder_conf
:
output_size
:
256
# dimension of attention
attention_heads
:
4
linear_units
:
2048
# the number of units of position-wise feed forward
num_blocks
:
12
# the number of encoder blocks
dropout_rate
:
0.1
positional_dropout_rate
:
0.1
attention_dropout_rate
:
0.0
input_layer
:
conv2d
# encoder input type, you can chose conv2d, conv2d6 and conv2d8
normalize_before
:
true
# decoder related
decoder
:
transformer
decoder_conf
:
attention_heads
:
4
linear_units
:
2048
num_blocks
:
6
dropout_rate
:
0.1
positional_dropout_rate
:
0.1
self_attention_dropout_rate
:
0.0
src_attention_dropout_rate
:
0.0
# hybrid CTC/attention
model_conf
:
asr_weight
:
0.0
ctc_weight
:
0.0
lsm_weight
:
0.1
# label smoothing option
length_normalized_loss
:
false
###########################################
# Training #
###########################################
n_epoch
:
40
accum_grad
:
2
global_grad_clip
:
5.0
optim
:
adam
optim_conf
:
lr
:
2.5
weight_decay
:
0.
scheduler
:
noam
scheduler_conf
:
warmup_steps
:
25000
lr_decay
:
1.0
log_interval
:
50
checkpoint
:
kbest_n
:
50
latest_n
:
5
\ No newline at end of file
编辑
预览
Markdown
is supported
0%
请重试
或
添加新附件
.
添加附件
取消
You are about to add
0
people
to the discussion. Proceed with caution.
先完成此消息的编辑!
取消
想要评论请
注册
或
登录