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

[pvt_v2_b0]训练精度不提升

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

• Software Environment / 软件环境 (Mandatory / 必填):
  -- MindSpore version (e.g., 1.7.0.Bxxx) :master_20240426230938_a87635b6f65bc6225173fec53883351ebf449b66
  -- Python version (e.g., Python 3.7.5) :Python 3.7.6
  -- OS platform and distribution (e.g., Linux Ubuntu 16.04):4.19.90-vhulk2211.3.0.h1543.eulerosv2r10.aarch64
  -- GCC/Compiler version (if compiled from source): gcc version 7.3.0 (GCC)
  run包Milan_C17/20240414
  mindspore包master_20240426230938_a87635b6f65bc6225173fec53883351ebf449b66

• 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 graph

Related testcase / 关联用例 (Mandatory / 必填)

test_ms_lab_pvt_v2_b0_acc3_ascend_train_8p_0008

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

1.get code from solution_test
2.cd solution_test/cases/02network/00cv/pvt_v2_b0/train/
3.pytest -s test_ms_lab_pvt_v2_b0_acc3_ascend_train_8p_0008.py
4.查看过训练精度是否正常

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

[pvt_v2_b0]训练精度正常提升

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

2024-04-27 05:10:13] mindcv.utils.callbacks INFO - Epoch: [34/34], batch: [1251/1251], loss: 9.847733, lr: 0.000996, time: 389.866129s
[2024-04-27 05:10:23] mindcv.utils.callbacks INFO - Validation Top_1_Accuracy: 0.1000%, Top_5_Accuracy: 0.5840%, time: 10.636981s
[2024-04-27 05:10:25] mindcv.utils.callbacks INFO - Saving model to ./ckpt/pvt_v2_b0-34_1251.ckpt
[2024-04-27 05:10:27] mindcv.utils.callbacks INFO - Total time since last epoch: 404.670255(train: 389.894190, val: 10.636981)s, ETA: 0.000000s
[2024-04-27 05:10:27] mindcv.utils.callbacks INFO - --------------------------------------------------------------------------------
[INFO] RUNTIME(66342,python):2024-04-27-05:13:32.464.990 [engine.cc:1709] 71397 ReportTimeoutProc: report timeout! streamId=1, taskId=357, execId=355, pendingNum=9, reportCount=427, parseTaskCount=427, msec=7991049, curSec=8175284
[INFO] RUNTIME(66342,python):2024-04-27-05:16:36.785.136 [engine.cc:1709] 71397 ReportTimeoutProc: report timeout! streamId=1, taskId=357, execId=355, pendingNum=9, reportCount=427, parseTaskCount=427, msec=7991049, curSec=8359604
[2024-04-27 05:17:30] mindcv.utils.callbacks INFO - Epoch: [35/34], batch: [1251/1251], loss: 10.055557, lr: 0.000996, time: 422.013266s
[2024-04-27 05:17:40] mindcv.utils.callbacks INFO - Validation Top_1_Accuracy: 0.1000%, Top_5_Accuracy: 0.5840%, time: 10.416377s
[2024-04-27 05:17:41] mindcv.utils.callbacks INFO - Saving model to ./ckpt/pvt_v2_b0-35_1251.ckpt
[2024-04-27 05:17:44] mindcv.utils.callbacks INFO - Total time since last epoch: 436.063551(train: 422.062537, val: 10.416377)s, ETA: -436.063551s
[2024-04-27 05:17:44] mindcv.utils.callbacks INFO - --------------------------------------------------------------------------------
[INFO] RUNTIME(66342,python):2024-04-27-05:20:48.684.866 [engine.cc:1709] 71397 ReportTimeoutProc: report timeout! streamId=1, taskId=531, execId=529, pendingNum=9, reportCount=587, parseTaskCount=587, msec=8427122, curSec=8611503
[INFO] RUNTIME(66342,python):2024-04-27-05:23:53.004.959 [engine.cc:1709] 71397 ReportTimeoutProc: report timeout! streamId=1, taskId=531, execId=529, pendingNum=9, reportCount=587, parseTaskCount=587, msec=8427122, curSec=8795824
[2024-04-27 05:24:37] mindcv.utils.callbacks INFO - Epoch: [36/34], batch: [1251/1251], loss: 10.048895, lr: 0.000996, time: 413.378937s
[2024-04-27 05:24:48] mindcv.utils.callbacks INFO - Validation Top_1_Accuracy: 0.1000%, Top_5_Accuracy: 0.5840%, time: 11.029083s
[2024-04-27 05:24:50] mindcv.utils.callbacks INFO - Saving model to ./ckpt/pvt_v2_b0-36_1251.ckpt
[2024-04-27 05:24:52] mindcv.utils.callbacks INFO - Total time since last epoch: 428.123073(train: 413.405590, val: 11.029083)s, ETA: -856.246147s
[2024-04-27 05:24:52] mindcv.utils.callbacks INFO - --------------------------------------------------------------------------------

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

可能与这个问题单有关联
https://e.gitee.com/mind_spore/issues/list?issue=I9HVXV

上次精度正常上升版本，也就是上一次ci下发执行该用例
run包Milan_C17/20240414
mindspore包master_20240417142516_74e1f3ea86a988bceb71bd7f02b674af4c97c89a/