Describe the current behavior / 问题描述 (Mandatory / 必填)

inceptionv3 imagenet2012数据集，修改了batch_size（与竞品对齐），在ASCEND环境，pynative模式，性能达不到规定要求（100%），在GPU环境，graph、pynative模式性能远远低于竞品性能
模型地址：https://gitee.com/mindspore/models/tree/master/official/cv/Inception/inceptionv3

Environment / 环境信息 (Mandatory / 必填)

• Hardware Environment(Ascend/GPU/CPU) / 硬件环境:

Please delete the backend not involved / 请删除不涉及的后端:
/device ascend/GPU/CPU/kirin/等其他芯片

• Software Environment / 软件环境 (Mandatory / 必填):
  -- MindSpore version (e.g., 1.7.0.Bxxx) :
  -- Python version (e.g., Python 3.7.5) :
  -- OS platform and distribution (e.g., Linux Ubuntu 16.04):
  -- GCC/Compiler version (if compiled from source):
  ascend(8卡)、GPU（1/8p）mindspore:2.0.0.20221220,commit_id:470b760e
  ascend(单卡)：mindspore:2.0.0.20221206,commit_id:0a257187

• Excute Mode / 执行模式 (Mandatory / 必填)(PyNative/Graph):

Please delete the mode not involved / 请删除不涉及的模式:
/mode pynative
/mode graph

Related testcase / 关联用例 (Mandatory / 必填)

用例仓地址：solution_test/cases/02network/00cv/inceptionv3/train
用例:
test_ms_inceptionv3_imagenet2012_train_check_fps_gpu_1p_0004.py
test_ms_inceptionv3_imagenet2012_train_check_loss_gpu_8p_0005.py
test_ms_inceptionv3_imagenet2012_train_check_fps_1p_0001.py
test_ms_inceptionv3_imagenet2012_train_check_loss_8p_0002.py

Steps to reproduce the issue / 重现步骤 (Mandatory / 必填)

1. get code from models
2. cd models/official/cv/Inception/inceptionv3
3. 跑ascend 8P用例时，需将 default_config.yaml中的batch_size修改为256，ascend(1p)\GPU(1/8p)则不需要修改batch_size
4. sh scripts/run_standalone_train_gpu.sh device_id data_root

Describe the expected behavior / 预期结果 (Mandatory / 必填)

ms/竞品性能比：ascend：100%，GPU：>=80%

Related log / screenshot / 日志 / 截图 (Mandatory / 必填)

ascend:
竞品：单卡 PT1.5+910 BATCH_SIZE:128
性能：769.965
mindspore：单卡
graph：1178
pynative：314.14
性能比：ms/pytorch=314.14/769.965=40.79%

竞品：8卡 PT1.5+910 BATCH_SIZE:2048
性能：5298.088
mindspore：8卡
graph：9997
pynative：3446.2
性能比：ms/pytorch=3446.2/5298.088=65.05%

GPU:
竞品：单卡 PT1.5+V100 BATCH_SIZE:128
性能：464
mindspore：单卡
graph：247
pynative：221.62
性能比：ms/pytorch=221.62/464=47.76%

竞品：8卡 PT1.5+V100 BATCH_SIZE:8*128
性能：2823
mindspore：8卡
graph：1851
pynative：1541.83
性能比：ms/pytorch=1541.83/2823=54.62%

Special notes for this issue/备注 (Optional / 选填)

走给安正气