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
PaddleClas
提交
890f7741
P
PaddleClas
项目概览
PaddlePaddle
/
PaddleClas
接近 2 年 前同步成功
通知
116
Star
4999
Fork
1114
代码
文件
提交
分支
Tags
贡献者
分支图
Diff
Issue
19
列表
看板
标记
里程碑
合并请求
6
Wiki
0
Wiki
分析
仓库
DevOps
项目成员
Pages
P
PaddleClas
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
19
Issue
19
列表
看板
标记
里程碑
合并请求
6
合并请求
6
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
提交
890f7741
编写于
4月 07, 2023
作者:
G
gaotingquan
提交者:
cuicheng01
5月 17, 2023
浏览文件
操作
浏览文件
下载
电子邮件补丁
差异文件
fix bs and unset update_freq to adapt to 8 gpus
上级
fc9c59c4
变更
9
隐藏空白更改
内联
并排
Showing
9 changed file
with
12 addition
and
21 deletion
+12
-21
ppcls/configs/ImageNet/MobileViTv3/MobileViTv3_S.yaml
ppcls/configs/ImageNet/MobileViTv3/MobileViTv3_S.yaml
+1
-2
ppcls/configs/ImageNet/MobileViTv3/MobileViTv3_S_L2.yaml
ppcls/configs/ImageNet/MobileViTv3/MobileViTv3_S_L2.yaml
+1
-2
ppcls/configs/ImageNet/MobileViTv3/MobileViTv3_XS.yaml
ppcls/configs/ImageNet/MobileViTv3/MobileViTv3_XS.yaml
+1
-2
ppcls/configs/ImageNet/MobileViTv3/MobileViTv3_XS_L2.yaml
ppcls/configs/ImageNet/MobileViTv3/MobileViTv3_XS_L2.yaml
+1
-2
ppcls/configs/ImageNet/MobileViTv3/MobileViTv3_XXS.yaml
ppcls/configs/ImageNet/MobileViTv3/MobileViTv3_XXS.yaml
+1
-2
ppcls/configs/ImageNet/MobileViTv3/MobileViTv3_XXS_L2.yaml
ppcls/configs/ImageNet/MobileViTv3/MobileViTv3_XXS_L2.yaml
+1
-2
ppcls/configs/ImageNet/MobileViTv3/MobileViTv3_x0_5.yaml
ppcls/configs/ImageNet/MobileViTv3/MobileViTv3_x0_5.yaml
+2
-3
ppcls/configs/ImageNet/MobileViTv3/MobileViTv3_x0_75.yaml
ppcls/configs/ImageNet/MobileViTv3/MobileViTv3_x0_75.yaml
+2
-3
ppcls/configs/ImageNet/MobileViTv3/MobileViTv3_x1_0.yaml
ppcls/configs/ImageNet/MobileViTv3/MobileViTv3_x1_0.yaml
+2
-3
未找到文件。
ppcls/configs/ImageNet/MobileViTv3/MobileViTv3_S.yaml
浏览文件 @
890f7741
...
...
@@ -14,7 +14,6 @@ Global:
image_shape
:
[
3
,
256
,
256
]
save_inference_dir
:
./inference
use_dali
:
False
update_freq
:
3
# for 4 gpus
# mixed precision training
AMP
:
...
...
@@ -85,7 +84,7 @@ DataLoader:
scales
:
[
256
,
160
,
192
,
224
,
288
,
320
]
# first_bs: batch size for the first image resolution in the scales list
# divide_factor: to ensure the width and height dimensions can be devided by downsampling multiple
first_bs
:
32
first_bs
:
48
divided_factor
:
32
is_training
:
True
loader
:
...
...
ppcls/configs/ImageNet/MobileViTv3/MobileViTv3_S_L2.yaml
浏览文件 @
890f7741
...
...
@@ -14,7 +14,6 @@ Global:
image_shape
:
[
3
,
256
,
256
]
save_inference_dir
:
./inference
use_dali
:
False
update_freq
:
3
# for 4 gpus
# mixed precision training
AMP
:
...
...
@@ -85,7 +84,7 @@ DataLoader:
scales
:
[
256
,
160
,
192
,
224
,
288
,
320
]
# first_bs: batch size for the first image resolution in the scales list
# divide_factor: to ensure the width and height dimensions can be devided by downsampling multiple
first_bs
:
32
first_bs
:
48
divided_factor
:
32
is_training
:
True
loader
:
...
...
ppcls/configs/ImageNet/MobileViTv3/MobileViTv3_XS.yaml
浏览文件 @
890f7741
...
...
@@ -14,7 +14,6 @@ Global:
image_shape
:
[
3
,
256
,
256
]
save_inference_dir
:
./inference
use_dali
:
False
update_freq
:
3
# for 4 gpus
# mixed precision training
AMP
:
...
...
@@ -85,7 +84,7 @@ DataLoader:
scales
:
[
256
,
160
,
192
,
224
,
288
,
320
]
# first_bs: batch size for the first image resolution in the scales list
# divide_factor: to ensure the width and height dimensions can be devided by downsampling multiple
first_bs
:
32
first_bs
:
48
divided_factor
:
32
is_training
:
True
loader
:
...
...
ppcls/configs/ImageNet/MobileViTv3/MobileViTv3_XS_L2.yaml
浏览文件 @
890f7741
...
...
@@ -14,7 +14,6 @@ Global:
image_shape
:
[
3
,
256
,
256
]
save_inference_dir
:
./inference
use_dali
:
False
update_freq
:
3
# for 4 gpus
# mixed precision training
AMP
:
...
...
@@ -85,7 +84,7 @@ DataLoader:
scales
:
[
256
,
160
,
192
,
224
,
288
,
320
]
# first_bs: batch size for the first image resolution in the scales list
# divide_factor: to ensure the width and height dimensions can be devided by downsampling multiple
first_bs
:
32
first_bs
:
48
divided_factor
:
32
is_training
:
True
loader
:
...
...
ppcls/configs/ImageNet/MobileViTv3/MobileViTv3_XXS.yaml
浏览文件 @
890f7741
...
...
@@ -14,7 +14,6 @@ Global:
image_shape
:
[
3
,
256
,
256
]
save_inference_dir
:
./inference
use_dali
:
False
update_freq
:
3
# for 4 gpus
# mixed precision training
AMP
:
...
...
@@ -85,7 +84,7 @@ DataLoader:
scales
:
[
256
,
160
,
192
,
224
,
288
,
320
]
# first_bs: batch size for the first image resolution in the scales list
# divide_factor: to ensure the width and height dimensions can be devided by downsampling multiple
first_bs
:
32
first_bs
:
48
divided_factor
:
32
is_training
:
True
loader
:
...
...
ppcls/configs/ImageNet/MobileViTv3/MobileViTv3_XXS_L2.yaml
浏览文件 @
890f7741
...
...
@@ -14,7 +14,6 @@ Global:
image_shape
:
[
3
,
256
,
256
]
save_inference_dir
:
./inference
use_dali
:
False
update_freq
:
3
# for 4 gpus
# mixed precision training
AMP
:
...
...
@@ -85,7 +84,7 @@ DataLoader:
scales
:
[
256
,
160
,
192
,
224
,
288
,
320
]
# first_bs: batch size for the first image resolution in the scales list
# divide_factor: to ensure the width and height dimensions can be devided by downsampling multiple
first_bs
:
32
first_bs
:
48
divided_factor
:
32
is_training
:
True
loader
:
...
...
ppcls/configs/ImageNet/MobileViTv3/MobileViTv3_x0_5.yaml
浏览文件 @
890f7741
...
...
@@ -14,7 +14,6 @@ Global:
image_shape
:
[
3
,
256
,
256
]
save_inference_dir
:
./inference
use_dali
:
False
update_freq
:
3
# for 4 gpus
# mixed precision training
AMP
:
...
...
@@ -52,7 +51,7 @@ Optimizer:
one_dim_param_no_weight_decay
:
True
lr
:
name
:
Cosine
learning_rate
:
0.002
# for total batch size 1020
learning_rate
:
0.002
# for total batch size 1020
by referring to official
eta_min
:
0.0002
warmup_epoch
:
16
# 20000 iterations
warmup_start_lr
:
1e-6
...
...
@@ -103,7 +102,7 @@ DataLoader:
prob
:
0.25
sampler
:
name
:
DistributedBatchSampler
batch_size
:
85
batch_size
:
128
drop_last
:
False
shuffle
:
True
loader
:
...
...
ppcls/configs/ImageNet/MobileViTv3/MobileViTv3_x0_75.yaml
浏览文件 @
890f7741
...
...
@@ -14,7 +14,6 @@ Global:
image_shape
:
[
3
,
256
,
256
]
save_inference_dir
:
./inference
use_dali
:
False
update_freq
:
3
# for 4 gpus
# mixed precision training
AMP
:
...
...
@@ -52,7 +51,7 @@ Optimizer:
one_dim_param_no_weight_decay
:
True
lr
:
name
:
Cosine
learning_rate
:
0.002
# for total batch size 1020
learning_rate
:
0.002
# for total batch size 1020
by referring to official
eta_min
:
0.0002
warmup_epoch
:
16
# 20000 iterations
warmup_start_lr
:
1e-6
...
...
@@ -103,7 +102,7 @@ DataLoader:
prob
:
0.25
sampler
:
name
:
DistributedBatchSampler
batch_size
:
85
batch_size
:
128
drop_last
:
False
shuffle
:
True
loader
:
...
...
ppcls/configs/ImageNet/MobileViTv3/MobileViTv3_x1_0.yaml
浏览文件 @
890f7741
...
...
@@ -14,7 +14,6 @@ Global:
image_shape
:
[
3
,
256
,
256
]
save_inference_dir
:
./inference
use_dali
:
False
update_freq
:
3
# for 4 gpus
# mixed precision training
AMP
:
...
...
@@ -52,7 +51,7 @@ Optimizer:
one_dim_param_no_weight_decay
:
True
lr
:
name
:
Cosine
learning_rate
:
0.002
# for total batch size 1020
learning_rate
:
0.002
# for total batch size 1020
by referring to official
eta_min
:
0.0002
warmup_epoch
:
16
# 20000 iterations
warmup_start_lr
:
1e-6
...
...
@@ -103,7 +102,7 @@ DataLoader:
prob
:
0.25
sampler
:
name
:
DistributedBatchSampler
batch_size
:
85
batch_size
:
128
drop_last
:
False
shuffle
:
True
loader
:
...
...
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