【众智】【东北大学】【ID2109】【CSGM】csgm精度性能达标 · Pull Request !335 · Ascend/ModelZoo-TensorFlow

csgm 裸机性能达标
modelart插件配置信息
![输入图片说明](https://images.gitee.com/uploads/images/2022/0518/192201_1f3a3937_9840899.png "屏幕截图.png")
执行结果打屏信息：
2022-03-21 23:33:35.866972: W /home/jenkins/agent/workspace/Compile_GraphEngine_Centos_X86/tensorflow/tf_adapter/util/ge_plugin.cc:124] [GePlugin] can not find Environment variable : JOB_ID
2022-03-21 23:33:39.807011: I /home/jenkins/agent/workspace/Compile_GraphEngine_Centos_X86/tensorflow/tf_adapter/kernels/geop_npu.cc:749] The model has been compiled on the Ascend AI processor, current graph id is:1
Initialized with PSNR: 18.26756789065104
2022-03-21 23:33:52.281454: I /home/jenkins/agent/workspace/Compile_GraphEngine_Centos_X86/tensorflow/tf_adapter/kernels/geop_npu.cc:749] The model has been compiled on the Ascend AI processor, current graph id is:11
step 12, loss = 103.19 (51654.0 examples/sec; 0.002 sec/batch)
step 12, loss = 103.04 (53065.6 examples/sec; 0.002 sec/batch)
step 12, loss = 98.33 (51807.1 examples/sec; 0.002 sec/batch)
step 12, loss = 100.44 (52187.4 examples/sec; 0.002 sec/batch)
step 12, loss = 98.23 (52911.6 examples/sec; 0.002 sec/batch)
step 12, loss = 103.22 (52083.7 examples/sec; 0.002 sec/batch)
step 12, loss = 102.44 (51974.0 examples/sec; 0.002 sec/batch)
step 12, loss = 103.07 (52389.5 examples/sec; 0.002 sec/batch)
step 12, loss = 99.82 (52435.4 examples/sec; 0.002 sec/batch)
step 12, loss = 100.67 (52402.6 examples/sec; 0.002 sec/batch)
step 12, loss = 103.34 (52851.6 examples/sec; 0.002 sec/batch)
step 12, loss = 104.92 (52632.8 examples/sec; 0.002 sec/batch)
step 12, loss = 99.91 (51539.7 examples/sec; 0.002 sec/batch)
step 12, loss = 101.96 (53410.2 examples/sec; 0.002 sec/batch)
step 12, loss = 103.25 (53166.5 examples/sec; 0.002 sec/batch)
step 12, loss = 107.41 (53335.5 examples/sec; 0.002 sec/batch)
step 12, loss = 106.33 (53546.6 examples/sec; 0.002 sec/batch)
step 12, loss = 104.44 (52291.5 examples/sec; 0.002 sec/batch)
step 12, loss = 97.90 (51609.5 examples/sec; 0.002 sec/batch)
step 12, loss = 101.81 (52298.1 examples/sec; 0.002 sec/batch)
step 12, loss = 104.01 (51590.5 examples/sec; 0.002 sec/batch)
step 12, loss = 98.72 (52032.1 examples/sec; 0.002 sec/batch)
step 12, loss = 98.53 (53200.2 examples/sec; 0.002 sec/batch)
step 12, loss = 98.74 (52317.6 examples/sec; 0.002 sec/batch)
step 12, loss = 105.42 (52232.9 examples/sec; 0.002 sec/batch)
step 12, loss = 103.77 (52665.8 examples/sec; 0.002 sec/batch)
step 12, loss = 102.57 (52396.1 examples/sec; 0.002 sec/batch)
step 12, loss = 99.36 (53234.0 examples/sec; 0.002 sec/batch)
step 12, loss = 101.95 (53615.0 examples/sec; 0.002 sec/batch)
step 12, loss = 105.18 (52422.2 examples/sec; 0.002 sec/batch)
step 12, loss = 102.93 (51704.9 examples/sec; 0.002 sec/batch)
step 12, loss = 100.61 (52369.9 examples/sec; 0.002 sec/batch)
step 12, loss = 106.17 (51225.0 examples/sec; 0.002 sec/batch)
step 12, loss = 102.04 (51813.5 examples/sec; 0.002 sec/batch)
step 12, loss = 107.66 (52369.9 examples/sec; 0.002 sec/batch)
step 12, loss = 109.57 (51673.1 examples/sec; 0.002 sec/batch)
step 12, loss = 104.66 (51388.2 examples/sec; 0.002 sec/batch)
step 12, loss = 101.40 (52514.1 examples/sec; 0.002 sec/batch)
step 12, loss = 99.98 (51337.9 examples/sec; 0.002 sec/batch)
step 12, loss = 103.62 (51916.1 examples/sec; 0.002 sec/batch)
step 12, loss = 101.46 (53105.9 examples/sec; 0.002 sec/batch)
step 12, loss = 104.52 (52599.7 examples/sec; 0.002 sec/batch)
step 12, loss = 99.36 (52396.1 examples/sec; 0.002 sec/batch)
step 12, loss = 95.70 (52884.9 examples/sec; 0.002 sec/batch)
step 12, loss = 103.42 (47148.2 examples/sec; 0.002 sec/batch)
step 12, loss = 102.05 (50889.4 examples/sec; 0.002 sec/batch)
step 12, loss = 104.18 (52441.9 examples/sec; 0.002 sec/batch)
step 12, loss = 102.44 (52109.6 examples/sec; 0.002 sec/batch)
step 12, loss = 101.26 (51546.1 examples/sec; 0.002 sec/batch)
step 12, loss = 102.05 (52468.2 examples/sec; 0.002 sec/batch)
step 12, loss = 102.58 (52051.4 examples/sec; 0.002 sec/batch)
step 12, loss = 99.98 (52540.4 examples/sec; 0.002 sec/batch)
step 12, loss = 103.54 (53553.4 examples/sec; 0.002 sec/batch)
step 12, loss = 102.53 (53342.3 examples/sec; 0.002 sec/batch)
step 12, loss = 104.03 (52402.6 examples/sec; 0.002 sec/batch)
step 12, loss = 99.35 (52805.0 examples/sec; 0.002 sec/batch)
step 12, loss = 104.60 (51903.3 examples/sec; 0.002 sec/batch)
step 12, loss = 104.71 (51896.9 examples/sec; 0.002 sec/batch)
step 12, loss = 103.18 (52945.0 examples/sec; 0.002 sec/batch)
step 12, loss = 98.93 (52019.1 examples/sec; 0.002 sec/batch)
step 12, loss = 101.61 (51858.4 examples/sec; 0.002 sec/batch)
step 12, loss = 105.64 (52187.4 examples/sec; 0.002 sec/batch)
step 12, loss = 98.17 (52070.8 examples/sec; 0.002 sec/batch)
step 12, loss = 101.07 (51698.6 examples/sec; 0.002 sec/batch)
step 12, loss = 96.97 (52858.3 examples/sec; 0.002 sec/batch)
step 12, loss = 101.28 (52045.0 examples/sec; 0.002 sec/batch)
执行任务OBS链接：
obs://dmsp-submit/npu/
（裸机代码）
数据集OBS链接：
obs://neu-error-log/CSGM性能不达标/mnist/
--------------------Ascend NPU INFO NOTICE----------------------
INFO,you used Ascend NPU,please check your result.
Final:
reconstruct_error：0.01100351967325015

Ascend/ModelZoo-TensorFlow .gitee-modal { width: 500px !important; }

搜索帮助

Ascend/ModelZoo-TensorFlow