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
09817fe8
P
PaddleClas
项目概览
PaddlePaddle
/
PaddleClas
大约 1 年 前同步成功
通知
115
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
09817fe8
编写于
5月 29, 2023
作者:
G
gaotingquan
提交者:
Tingquan Gao
5月 29, 2023
浏览文件
操作
浏览文件
下载
电子邮件补丁
差异文件
complete amp args
上级
b3f7e3b9
变更
34
隐藏空白更改
内联
并排
Showing
34 changed file
with
102 addition
and
0 deletion
+102
-0
ppcls/configs/ImageNet/Distillation/PPLCNet_x2_5_dml.yaml
ppcls/configs/ImageNet/Distillation/PPLCNet_x2_5_dml.yaml
+3
-0
ppcls/configs/ImageNet/Distillation/PPLCNet_x2_5_ssld.yaml
ppcls/configs/ImageNet/Distillation/PPLCNet_x2_5_ssld.yaml
+3
-0
ppcls/configs/ImageNet/Distillation/PPLCNet_x2_5_udml.yaml
ppcls/configs/ImageNet/Distillation/PPLCNet_x2_5_udml.yaml
+3
-0
ppcls/configs/ImageNet/Distillation/res2net200_vd_distill_pphgnet_base.yaml
...eNet/Distillation/res2net200_vd_distill_pphgnet_base.yaml
+3
-0
ppcls/configs/ImageNet/EfficientNetV2/EfficientNetV2_S.yaml
ppcls/configs/ImageNet/EfficientNetV2/EfficientNetV2_S.yaml
+3
-0
ppcls/configs/ImageNet/MobileNetV3/MobileNetV3_small_x1_0_ampo2_ultra.yaml
...geNet/MobileNetV3/MobileNetV3_small_x1_0_ampo2_ultra.yaml
+3
-0
ppcls/configs/ImageNet/MobileViTV2/MobileViTV2_x0_5.yaml
ppcls/configs/ImageNet/MobileViTV2/MobileViTV2_x0_5.yaml
+3
-0
ppcls/configs/ImageNet/MobileViTV2/MobileViTV2_x1_0.yaml
ppcls/configs/ImageNet/MobileViTV2/MobileViTV2_x1_0.yaml
+3
-0
ppcls/configs/ImageNet/MobileViTV2/MobileViTV2_x1_5.yaml
ppcls/configs/ImageNet/MobileViTV2/MobileViTV2_x1_5.yaml
+3
-0
ppcls/configs/ImageNet/MobileViTV2/MobileViTV2_x2_0.yaml
ppcls/configs/ImageNet/MobileViTV2/MobileViTV2_x2_0.yaml
+3
-0
ppcls/configs/ImageNet/MobileViTV3/MobileViTV3_S.yaml
ppcls/configs/ImageNet/MobileViTV3/MobileViTV3_S.yaml
+3
-0
ppcls/configs/ImageNet/MobileViTV3/MobileViTV3_S_L2.yaml
ppcls/configs/ImageNet/MobileViTV3/MobileViTV3_S_L2.yaml
+3
-0
ppcls/configs/ImageNet/MobileViTV3/MobileViTV3_XS.yaml
ppcls/configs/ImageNet/MobileViTV3/MobileViTV3_XS.yaml
+3
-0
ppcls/configs/ImageNet/MobileViTV3/MobileViTV3_XS_L2.yaml
ppcls/configs/ImageNet/MobileViTV3/MobileViTV3_XS_L2.yaml
+3
-0
ppcls/configs/ImageNet/MobileViTV3/MobileViTV3_XXS.yaml
ppcls/configs/ImageNet/MobileViTV3/MobileViTV3_XXS.yaml
+3
-0
ppcls/configs/ImageNet/MobileViTV3/MobileViTV3_XXS_L2.yaml
ppcls/configs/ImageNet/MobileViTV3/MobileViTV3_XXS_L2.yaml
+3
-0
ppcls/configs/ImageNet/MobileViTV3/MobileViTV3_x0_5.yaml
ppcls/configs/ImageNet/MobileViTV3/MobileViTV3_x0_5.yaml
+3
-0
ppcls/configs/ImageNet/MobileViTV3/MobileViTV3_x0_75.yaml
ppcls/configs/ImageNet/MobileViTV3/MobileViTV3_x0_75.yaml
+3
-0
ppcls/configs/ImageNet/MobileViTV3/MobileViTV3_x1_0.yaml
ppcls/configs/ImageNet/MobileViTV3/MobileViTV3_x1_0.yaml
+3
-0
ppcls/configs/ImageNet/NextViT/NextViT_base_224.yaml
ppcls/configs/ImageNet/NextViT/NextViT_base_224.yaml
+3
-0
ppcls/configs/ImageNet/NextViT/NextViT_base_384.yaml
ppcls/configs/ImageNet/NextViT/NextViT_base_384.yaml
+3
-0
ppcls/configs/ImageNet/NextViT/NextViT_large_224.yaml
ppcls/configs/ImageNet/NextViT/NextViT_large_224.yaml
+3
-0
ppcls/configs/ImageNet/NextViT/NextViT_large_384.yaml
ppcls/configs/ImageNet/NextViT/NextViT_large_384.yaml
+3
-0
ppcls/configs/ImageNet/NextViT/NextViT_small_224.yaml
ppcls/configs/ImageNet/NextViT/NextViT_small_224.yaml
+3
-0
ppcls/configs/ImageNet/NextViT/NextViT_small_384.yaml
ppcls/configs/ImageNet/NextViT/NextViT_small_384.yaml
+3
-0
ppcls/configs/ImageNet/PPHGNet/PPHGNet_base.yaml
ppcls/configs/ImageNet/PPHGNet/PPHGNet_base.yaml
+3
-0
ppcls/configs/ImageNet/PPHGNet/PPHGNet_small.yaml
ppcls/configs/ImageNet/PPHGNet/PPHGNet_small.yaml
+3
-0
ppcls/configs/ImageNet/PPHGNet/PPHGNet_tiny.yaml
ppcls/configs/ImageNet/PPHGNet/PPHGNet_tiny.yaml
+3
-0
ppcls/configs/ImageNet/PPLCNet/PPLCNet_x1_0_ampo2_ultra.yaml
ppcls/configs/ImageNet/PPLCNet/PPLCNet_x1_0_ampo2_ultra.yaml
+3
-0
ppcls/configs/ImageNet/ResNet/ResNet50_amp_O1.yaml
ppcls/configs/ImageNet/ResNet/ResNet50_amp_O1.yaml
+3
-0
ppcls/configs/ImageNet/ResNet/ResNet50_amp_O1_ultra.yaml
ppcls/configs/ImageNet/ResNet/ResNet50_amp_O1_ultra.yaml
+3
-0
ppcls/configs/ImageNet/ResNet/ResNet50_amp_O2_ultra.yaml
ppcls/configs/ImageNet/ResNet/ResNet50_amp_O2_ultra.yaml
+3
-0
ppcls/configs/ImageNet/ResNet/ResNet50_ampo2_ultra.yaml
ppcls/configs/ImageNet/ResNet/ResNet50_ampo2_ultra.yaml
+3
-0
ppcls/configs/ImageNet/SENet/SE_ResNeXt101_32x4d_amp_O2_ultra.yaml
...figs/ImageNet/SENet/SE_ResNeXt101_32x4d_amp_O2_ultra.yaml
+3
-0
未找到文件。
ppcls/configs/ImageNet/Distillation/PPLCNet_x2_5_dml.yaml
浏览文件 @
09817fe8
...
@@ -15,8 +15,11 @@ Global:
...
@@ -15,8 +15,11 @@ Global:
save_inference_dir
:
./inference
save_inference_dir
:
./inference
AMP
:
AMP
:
use_amp
:
True
use_fp16_test
:
False
scale_loss
:
128.0
scale_loss
:
128.0
use_dynamic_loss_scaling
:
True
use_dynamic_loss_scaling
:
True
use_promote
:
False
# O1: mixed fp16, O2: pure fp16
# O1: mixed fp16, O2: pure fp16
level
:
O1
level
:
O1
...
...
ppcls/configs/ImageNet/Distillation/PPLCNet_x2_5_ssld.yaml
浏览文件 @
09817fe8
...
@@ -16,8 +16,11 @@ Global:
...
@@ -16,8 +16,11 @@ Global:
to_static
:
True
to_static
:
True
AMP
:
AMP
:
use_amp
:
True
use_fp16_test
:
False
scale_loss
:
128.0
scale_loss
:
128.0
use_dynamic_loss_scaling
:
True
use_dynamic_loss_scaling
:
True
use_promote
:
False
# O1: mixed fp16, O2: pure fp16
# O1: mixed fp16, O2: pure fp16
level
:
O1
level
:
O1
...
...
ppcls/configs/ImageNet/Distillation/PPLCNet_x2_5_udml.yaml
浏览文件 @
09817fe8
...
@@ -15,8 +15,11 @@ Global:
...
@@ -15,8 +15,11 @@ Global:
save_inference_dir
:
./inference
save_inference_dir
:
./inference
AMP
:
AMP
:
use_amp
:
True
use_fp16_test
:
False
scale_loss
:
128.0
scale_loss
:
128.0
use_dynamic_loss_scaling
:
True
use_dynamic_loss_scaling
:
True
use_promote
:
False
# O1: mixed fp16, O2: pure fp16
# O1: mixed fp16, O2: pure fp16
level
:
O1
level
:
O1
...
...
ppcls/configs/ImageNet/Distillation/res2net200_vd_distill_pphgnet_base.yaml
浏览文件 @
09817fe8
...
@@ -17,8 +17,11 @@ Global:
...
@@ -17,8 +17,11 @@ Global:
# mixed precision training
# mixed precision training
AMP
:
AMP
:
use_amp
:
True
use_fp16_test
:
False
scale_loss
:
128.0
scale_loss
:
128.0
use_dynamic_loss_scaling
:
True
use_dynamic_loss_scaling
:
True
use_promote
:
False
# O1: mixed fp16, O2: pure fp16
# O1: mixed fp16, O2: pure fp16
level
:
O1
level
:
O1
...
...
ppcls/configs/ImageNet/EfficientNetV2/EfficientNetV2_S.yaml
浏览文件 @
09817fe8
...
@@ -16,8 +16,11 @@ Global:
...
@@ -16,8 +16,11 @@ Global:
save_inference_dir
:
./inference
save_inference_dir
:
./inference
AMP
:
AMP
:
use_amp
:
True
use_fp16_test
:
False
scale_loss
:
65536
scale_loss
:
65536
use_dynamic_loss_scaling
:
True
use_dynamic_loss_scaling
:
True
use_promote
:
False
# O1: mixed fp16, O2: pure fp16
# O1: mixed fp16, O2: pure fp16
level
:
O1
level
:
O1
...
...
ppcls/configs/ImageNet/MobileNetV3/MobileNetV3_small_x1_0_ampo2_ultra.yaml
浏览文件 @
09817fe8
...
@@ -17,8 +17,11 @@ Global:
...
@@ -17,8 +17,11 @@ Global:
# mixed precision training
# mixed precision training
AMP
:
AMP
:
use_amp
:
True
use_fp16_test
:
False
scale_loss
:
128.0
scale_loss
:
128.0
use_dynamic_loss_scaling
:
True
use_dynamic_loss_scaling
:
True
use_promote
:
False
# O1: mixed fp16, O2: pure fp16
# O1: mixed fp16, O2: pure fp16
level
:
O2
level
:
O2
...
...
ppcls/configs/ImageNet/MobileViTV2/MobileViTV2_x0_5.yaml
浏览文件 @
09817fe8
...
@@ -17,8 +17,11 @@ Global:
...
@@ -17,8 +17,11 @@ Global:
# mixed precision training
# mixed precision training
AMP
:
AMP
:
use_amp
:
True
use_fp16_test
:
False
scale_loss
:
65536
scale_loss
:
65536
use_dynamic_loss_scaling
:
True
use_dynamic_loss_scaling
:
True
use_promote
:
False
# O1: mixed fp16, O2: pure fp16
# O1: mixed fp16, O2: pure fp16
level
:
O1
level
:
O1
...
...
ppcls/configs/ImageNet/MobileViTV2/MobileViTV2_x1_0.yaml
浏览文件 @
09817fe8
...
@@ -17,8 +17,11 @@ Global:
...
@@ -17,8 +17,11 @@ Global:
# mixed precision training
# mixed precision training
AMP
:
AMP
:
use_amp
:
True
use_fp16_test
:
False
scale_loss
:
65536
scale_loss
:
65536
use_dynamic_loss_scaling
:
True
use_dynamic_loss_scaling
:
True
use_promote
:
False
# O1: mixed fp16, O2: pure fp16
# O1: mixed fp16, O2: pure fp16
level
:
O1
level
:
O1
...
...
ppcls/configs/ImageNet/MobileViTV2/MobileViTV2_x1_5.yaml
浏览文件 @
09817fe8
...
@@ -17,8 +17,11 @@ Global:
...
@@ -17,8 +17,11 @@ Global:
# mixed precision training
# mixed precision training
AMP
:
AMP
:
use_amp
:
True
use_fp16_test
:
False
scale_loss
:
65536
scale_loss
:
65536
use_dynamic_loss_scaling
:
True
use_dynamic_loss_scaling
:
True
use_promote
:
False
# O1: mixed fp16, O2: pure fp16
# O1: mixed fp16, O2: pure fp16
level
:
O1
level
:
O1
...
...
ppcls/configs/ImageNet/MobileViTV2/MobileViTV2_x2_0.yaml
浏览文件 @
09817fe8
...
@@ -17,8 +17,11 @@ Global:
...
@@ -17,8 +17,11 @@ Global:
# mixed precision training
# mixed precision training
AMP
:
AMP
:
use_amp
:
True
use_fp16_test
:
False
scale_loss
:
65536
scale_loss
:
65536
use_dynamic_loss_scaling
:
True
use_dynamic_loss_scaling
:
True
use_promote
:
False
# O1: mixed fp16, O2: pure fp16
# O1: mixed fp16, O2: pure fp16
level
:
O1
level
:
O1
...
...
ppcls/configs/ImageNet/MobileViTV3/MobileViTV3_S.yaml
浏览文件 @
09817fe8
...
@@ -17,8 +17,11 @@ Global:
...
@@ -17,8 +17,11 @@ Global:
# mixed precision training
# mixed precision training
AMP
:
AMP
:
use_amp
:
True
use_fp16_test
:
False
scale_loss
:
65536
scale_loss
:
65536
use_dynamic_loss_scaling
:
True
use_dynamic_loss_scaling
:
True
use_promote
:
False
# O1: mixed fp16, O2: pure fp16
# O1: mixed fp16, O2: pure fp16
level
:
O1
level
:
O1
...
...
ppcls/configs/ImageNet/MobileViTV3/MobileViTV3_S_L2.yaml
浏览文件 @
09817fe8
...
@@ -17,8 +17,11 @@ Global:
...
@@ -17,8 +17,11 @@ Global:
# mixed precision training
# mixed precision training
AMP
:
AMP
:
use_amp
:
True
use_fp16_test
:
False
scale_loss
:
65536
scale_loss
:
65536
use_dynamic_loss_scaling
:
True
use_dynamic_loss_scaling
:
True
use_promote
:
False
# O1: mixed fp16, O2: pure fp16
# O1: mixed fp16, O2: pure fp16
level
:
O1
level
:
O1
...
...
ppcls/configs/ImageNet/MobileViTV3/MobileViTV3_XS.yaml
浏览文件 @
09817fe8
...
@@ -17,8 +17,11 @@ Global:
...
@@ -17,8 +17,11 @@ Global:
# mixed precision training
# mixed precision training
AMP
:
AMP
:
use_amp
:
True
use_fp16_test
:
False
scale_loss
:
65536
scale_loss
:
65536
use_dynamic_loss_scaling
:
True
use_dynamic_loss_scaling
:
True
use_promote
:
False
# O1: mixed fp16, O2: pure fp16
# O1: mixed fp16, O2: pure fp16
level
:
O1
level
:
O1
...
...
ppcls/configs/ImageNet/MobileViTV3/MobileViTV3_XS_L2.yaml
浏览文件 @
09817fe8
...
@@ -17,8 +17,11 @@ Global:
...
@@ -17,8 +17,11 @@ Global:
# mixed precision training
# mixed precision training
AMP
:
AMP
:
use_amp
:
True
use_fp16_test
:
False
scale_loss
:
65536
scale_loss
:
65536
use_dynamic_loss_scaling
:
True
use_dynamic_loss_scaling
:
True
use_promote
:
False
# O1: mixed fp16, O2: pure fp16
# O1: mixed fp16, O2: pure fp16
level
:
O1
level
:
O1
...
...
ppcls/configs/ImageNet/MobileViTV3/MobileViTV3_XXS.yaml
浏览文件 @
09817fe8
...
@@ -17,8 +17,11 @@ Global:
...
@@ -17,8 +17,11 @@ Global:
# mixed precision training
# mixed precision training
AMP
:
AMP
:
use_amp
:
True
use_fp16_test
:
False
scale_loss
:
65536
scale_loss
:
65536
use_dynamic_loss_scaling
:
True
use_dynamic_loss_scaling
:
True
use_promote
:
False
# O1: mixed fp16, O2: pure fp16
# O1: mixed fp16, O2: pure fp16
level
:
O1
level
:
O1
...
...
ppcls/configs/ImageNet/MobileViTV3/MobileViTV3_XXS_L2.yaml
浏览文件 @
09817fe8
...
@@ -17,8 +17,11 @@ Global:
...
@@ -17,8 +17,11 @@ Global:
# mixed precision training
# mixed precision training
AMP
:
AMP
:
use_amp
:
True
use_fp16_test
:
False
scale_loss
:
65536
scale_loss
:
65536
use_dynamic_loss_scaling
:
True
use_dynamic_loss_scaling
:
True
use_promote
:
False
# O1: mixed fp16, O2: pure fp16
# O1: mixed fp16, O2: pure fp16
level
:
O1
level
:
O1
...
...
ppcls/configs/ImageNet/MobileViTV3/MobileViTV3_x0_5.yaml
浏览文件 @
09817fe8
...
@@ -17,8 +17,11 @@ Global:
...
@@ -17,8 +17,11 @@ Global:
# mixed precision training
# mixed precision training
AMP
:
AMP
:
use_amp
:
True
use_fp16_test
:
False
scale_loss
:
65536
scale_loss
:
65536
use_dynamic_loss_scaling
:
True
use_dynamic_loss_scaling
:
True
use_promote
:
False
# O1: mixed fp16, O2: pure fp16
# O1: mixed fp16, O2: pure fp16
level
:
O1
level
:
O1
...
...
ppcls/configs/ImageNet/MobileViTV3/MobileViTV3_x0_75.yaml
浏览文件 @
09817fe8
...
@@ -17,8 +17,11 @@ Global:
...
@@ -17,8 +17,11 @@ Global:
# mixed precision training
# mixed precision training
AMP
:
AMP
:
use_amp
:
True
use_fp16_test
:
False
scale_loss
:
65536
scale_loss
:
65536
use_dynamic_loss_scaling
:
True
use_dynamic_loss_scaling
:
True
use_promote
:
False
# O1: mixed fp16, O2: pure fp16
# O1: mixed fp16, O2: pure fp16
level
:
O1
level
:
O1
...
...
ppcls/configs/ImageNet/MobileViTV3/MobileViTV3_x1_0.yaml
浏览文件 @
09817fe8
...
@@ -17,8 +17,11 @@ Global:
...
@@ -17,8 +17,11 @@ Global:
# mixed precision training
# mixed precision training
AMP
:
AMP
:
use_amp
:
True
use_fp16_test
:
False
scale_loss
:
65536
scale_loss
:
65536
use_dynamic_loss_scaling
:
True
use_dynamic_loss_scaling
:
True
use_promote
:
False
# O1: mixed fp16, O2: pure fp16
# O1: mixed fp16, O2: pure fp16
level
:
O1
level
:
O1
...
...
ppcls/configs/ImageNet/NextViT/NextViT_base_224.yaml
浏览文件 @
09817fe8
...
@@ -19,8 +19,11 @@ Global:
...
@@ -19,8 +19,11 @@ Global:
# mixed precision training
# mixed precision training
AMP
:
AMP
:
use_amp
:
True
use_fp16_test
:
False
scale_loss
:
128.0
scale_loss
:
128.0
use_dynamic_loss_scaling
:
True
use_dynamic_loss_scaling
:
True
use_promote
:
False
# O1: mixed fp16, O2: pure fp16
# O1: mixed fp16, O2: pure fp16
level
:
O1
level
:
O1
...
...
ppcls/configs/ImageNet/NextViT/NextViT_base_384.yaml
浏览文件 @
09817fe8
...
@@ -19,8 +19,11 @@ Global:
...
@@ -19,8 +19,11 @@ Global:
# mixed precision training
# mixed precision training
AMP
:
AMP
:
use_amp
:
True
use_fp16_test
:
False
scale_loss
:
128.0
scale_loss
:
128.0
use_dynamic_loss_scaling
:
True
use_dynamic_loss_scaling
:
True
use_promote
:
False
# O1: mixed fp16, O2: pure fp16
# O1: mixed fp16, O2: pure fp16
level
:
O1
level
:
O1
...
...
ppcls/configs/ImageNet/NextViT/NextViT_large_224.yaml
浏览文件 @
09817fe8
...
@@ -19,8 +19,11 @@ Global:
...
@@ -19,8 +19,11 @@ Global:
# mixed precision training
# mixed precision training
AMP
:
AMP
:
use_amp
:
True
use_fp16_test
:
False
scale_loss
:
128.0
scale_loss
:
128.0
use_dynamic_loss_scaling
:
True
use_dynamic_loss_scaling
:
True
use_promote
:
False
# O1: mixed fp16, O2: pure fp16
# O1: mixed fp16, O2: pure fp16
level
:
O1
level
:
O1
...
...
ppcls/configs/ImageNet/NextViT/NextViT_large_384.yaml
浏览文件 @
09817fe8
...
@@ -19,8 +19,11 @@ Global:
...
@@ -19,8 +19,11 @@ Global:
# mixed precision training
# mixed precision training
AMP
:
AMP
:
use_amp
:
True
use_fp16_test
:
False
scale_loss
:
128.0
scale_loss
:
128.0
use_dynamic_loss_scaling
:
True
use_dynamic_loss_scaling
:
True
use_promote
:
False
# O1: mixed fp16, O2: pure fp16
# O1: mixed fp16, O2: pure fp16
level
:
O1
level
:
O1
...
...
ppcls/configs/ImageNet/NextViT/NextViT_small_224.yaml
浏览文件 @
09817fe8
...
@@ -19,8 +19,11 @@ Global:
...
@@ -19,8 +19,11 @@ Global:
# mixed precision training
# mixed precision training
AMP
:
AMP
:
use_amp
:
True
use_fp16_test
:
False
scale_loss
:
128.0
scale_loss
:
128.0
use_dynamic_loss_scaling
:
True
use_dynamic_loss_scaling
:
True
use_promote
:
False
# O1: mixed fp16, O2: pure fp16
# O1: mixed fp16, O2: pure fp16
level
:
O1
level
:
O1
...
...
ppcls/configs/ImageNet/NextViT/NextViT_small_384.yaml
浏览文件 @
09817fe8
...
@@ -19,8 +19,11 @@ Global:
...
@@ -19,8 +19,11 @@ Global:
# mixed precision training
# mixed precision training
AMP
:
AMP
:
use_amp
:
True
use_fp16_test
:
False
scale_loss
:
128.0
scale_loss
:
128.0
use_dynamic_loss_scaling
:
True
use_dynamic_loss_scaling
:
True
use_promote
:
False
# O1: mixed fp16, O2: pure fp16
# O1: mixed fp16, O2: pure fp16
level
:
O1
level
:
O1
...
...
ppcls/configs/ImageNet/PPHGNet/PPHGNet_base.yaml
浏览文件 @
09817fe8
...
@@ -19,8 +19,11 @@ Global:
...
@@ -19,8 +19,11 @@ Global:
# mixed precision training
# mixed precision training
AMP
:
AMP
:
use_amp
:
True
use_fp16_test
:
False
scale_loss
:
128.0
scale_loss
:
128.0
use_dynamic_loss_scaling
:
True
use_dynamic_loss_scaling
:
True
use_promote
:
False
# O1: mixed fp16, O2: pure fp16
# O1: mixed fp16, O2: pure fp16
level
:
O1
level
:
O1
...
...
ppcls/configs/ImageNet/PPHGNet/PPHGNet_small.yaml
浏览文件 @
09817fe8
...
@@ -19,8 +19,11 @@ Global:
...
@@ -19,8 +19,11 @@ Global:
# mixed precision training
# mixed precision training
AMP
:
AMP
:
use_amp
:
True
use_fp16_test
:
False
scale_loss
:
128.0
scale_loss
:
128.0
use_dynamic_loss_scaling
:
True
use_dynamic_loss_scaling
:
True
use_promote
:
False
# O1: mixed fp16, O2: pure fp16
# O1: mixed fp16, O2: pure fp16
level
:
O1
level
:
O1
...
...
ppcls/configs/ImageNet/PPHGNet/PPHGNet_tiny.yaml
浏览文件 @
09817fe8
...
@@ -19,8 +19,11 @@ Global:
...
@@ -19,8 +19,11 @@ Global:
# mixed precision training
# mixed precision training
AMP
:
AMP
:
use_amp
:
True
use_fp16_test
:
False
scale_loss
:
128.0
scale_loss
:
128.0
use_dynamic_loss_scaling
:
True
use_dynamic_loss_scaling
:
True
use_promote
:
False
# O1: mixed fp16, O2: pure fp16
# O1: mixed fp16, O2: pure fp16
level
:
O1
level
:
O1
...
...
ppcls/configs/ImageNet/PPLCNet/PPLCNet_x1_0_ampo2_ultra.yaml
浏览文件 @
09817fe8
...
@@ -17,8 +17,11 @@ Global:
...
@@ -17,8 +17,11 @@ Global:
# mixed precision training
# mixed precision training
AMP
:
AMP
:
use_amp
:
True
use_fp16_test
:
False
scale_loss
:
128.0
scale_loss
:
128.0
use_dynamic_loss_scaling
:
True
use_dynamic_loss_scaling
:
True
use_promote
:
False
# O1: mixed fp16, O2: pure fp16
# O1: mixed fp16, O2: pure fp16
level
:
O2
level
:
O2
...
...
ppcls/configs/ImageNet/ResNet/ResNet50_amp_O1.yaml
浏览文件 @
09817fe8
...
@@ -19,8 +19,11 @@ Global:
...
@@ -19,8 +19,11 @@ Global:
# mixed precision training
# mixed precision training
AMP
:
AMP
:
use_amp
:
True
use_fp16_test
:
False
scale_loss
:
128.0
scale_loss
:
128.0
use_dynamic_loss_scaling
:
True
use_dynamic_loss_scaling
:
True
use_promote
:
False
# O1: mixed fp16, O2: pure fp16
# O1: mixed fp16, O2: pure fp16
level
:
O1
level
:
O1
...
...
ppcls/configs/ImageNet/ResNet/ResNet50_amp_O1_ultra.yaml
浏览文件 @
09817fe8
...
@@ -20,8 +20,11 @@ Global:
...
@@ -20,8 +20,11 @@ Global:
# mixed precision training
# mixed precision training
AMP
:
AMP
:
use_amp
:
True
use_fp16_test
:
False
scale_loss
:
128.0
scale_loss
:
128.0
use_dynamic_loss_scaling
:
True
use_dynamic_loss_scaling
:
True
use_promote
:
False
# O1: mixed fp16, O2: pure fp16
# O1: mixed fp16, O2: pure fp16
level
:
O1
level
:
O1
...
...
ppcls/configs/ImageNet/ResNet/ResNet50_amp_O2_ultra.yaml
浏览文件 @
09817fe8
...
@@ -20,8 +20,11 @@ Global:
...
@@ -20,8 +20,11 @@ Global:
# mixed precision training
# mixed precision training
AMP
:
AMP
:
use_amp
:
True
use_fp16_test
:
False
scale_loss
:
128.0
scale_loss
:
128.0
use_dynamic_loss_scaling
:
True
use_dynamic_loss_scaling
:
True
use_promote
:
False
# O1: mixed fp16, O2: pure fp16
# O1: mixed fp16, O2: pure fp16
level
:
O2
level
:
O2
...
...
ppcls/configs/ImageNet/ResNet/ResNet50_ampo2_ultra.yaml
浏览文件 @
09817fe8
...
@@ -19,8 +19,11 @@ Global:
...
@@ -19,8 +19,11 @@ Global:
# mixed precision training
# mixed precision training
AMP
:
AMP
:
use_amp
:
True
use_fp16_test
:
False
scale_loss
:
128.0
scale_loss
:
128.0
use_dynamic_loss_scaling
:
True
use_dynamic_loss_scaling
:
True
use_promote
:
False
# O1: mixed fp16, O2: pure fp16
# O1: mixed fp16, O2: pure fp16
level
:
O2
level
:
O2
...
...
ppcls/configs/ImageNet/SENet/SE_ResNeXt101_32x4d_amp_O2_ultra.yaml
浏览文件 @
09817fe8
...
@@ -17,8 +17,11 @@ Global:
...
@@ -17,8 +17,11 @@ Global:
# mixed precision training
# mixed precision training
AMP
:
AMP
:
use_amp
:
True
use_fp16_test
:
False
scale_loss
:
128.0
scale_loss
:
128.0
use_dynamic_loss_scaling
:
True
use_dynamic_loss_scaling
:
True
use_promote
:
False
# O1: mixed fp16, O2: pure fp16
# O1: mixed fp16, O2: pure fp16
level
:
O2
level
:
O2
...
...
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