singa-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From GitBox <...@apache.org>
Subject [GitHub] [singa] chrishkchris commented on issue #651: [WIP ]Simply example APIs
Date Sun, 05 Apr 2020 14:18:55 GMT
chrishkchris commented on issue #651: [WIP ]Simply example APIs
URL: https://github.com/apache/singa/pull/651#issuecomment-609423668
 
 
   I followed the new design of API, now can run (i) cnn+mnist, (ii) resnet+cifar10 
   ```
   root@877a3759b148:~/dcsysh/singa/examples/autograd# python3 train_multiprocess.py cnn mnist
--lr 0.01
   Starting Epoch 0:
   Training loss = 625.124268, training accuracy = 0.777227
   Evaluation accuracy = 0.941206, Elapsed Time = 1.721721s
   Starting Epoch 1:
   Training loss = 235.566132, training accuracy = 0.920757
   Evaluation accuracy = 0.945413, Elapsed Time = 1.685381s
   Starting Epoch 2:
   Training loss = 171.600082, training accuracy = 0.942258
   Evaluation accuracy = 0.969852, Elapsed Time = 1.703893s
   Starting Epoch 3:
   Training loss = 139.339203, training accuracy = 0.953476
   Evaluation accuracy = 0.972857, Elapsed Time = 1.695866s
   Starting Epoch 4:
   Training loss = 120.622467, training accuracy = 0.959118
   Evaluation accuracy = 0.971254, Elapsed Time = 1.703533s
   Starting Epoch 5:
   Training loss = 105.304459, training accuracy = 0.964777
   Evaluation accuracy = 0.978466, Elapsed Time = 1.702236s
   Starting Epoch 6:
   Training loss = 99.502411, training accuracy = 0.966964
   Evaluation accuracy = 0.975761, Elapsed Time = 1.694606s
   Starting Epoch 7:
   Training loss = 88.000076, training accuracy = 0.969985
   Evaluation accuracy = 0.977364, Elapsed Time = 1.700657s
   Starting Epoch 8:
   Training loss = 85.234161, training accuracy = 0.971004
   Evaluation accuracy = 0.976262, Elapsed Time = 1.690182s
   Starting Epoch 9:
   Training loss = 79.724716, training accuracy = 0.973591
   Evaluation accuracy = 0.983073, Elapsed Time = 1.701022s
   
   root@877a3759b148:~/dcsysh/singa/examples/autograd# python3 train_multiprocess.py resnet
cifar10 --lr 0.01 --bs 32
   Check the shape of dataset:
   Starting Epoch 0:
   Training loss = 3112.784912, training accuracy = 0.317702
   Evaluation accuracy = 0.464543, Elapsed Time = 151.415575s
   Starting Epoch 1:
   Training loss = 2140.442383, training accuracy = 0.500840
   Evaluation accuracy = 0.568710, Elapsed Time = 156.812310s
   Starting Epoch 2:
   Training loss = 1680.163208, training accuracy = 0.616997
   Evaluation accuracy = 0.681891, Elapsed Time = 157.775971s
   Starting Epoch 3:
   Training loss = 1395.033936, training accuracy = 0.688500
   Evaluation accuracy = 0.711939, Elapsed Time = 157.391661s
   Starting Epoch 4:
   Training loss = 1174.264160, training accuracy = 0.738916
   Evaluation accuracy = 0.756711, Elapsed Time = 157.423720s
   Starting Epoch 5:
   Training loss = 1028.127197, training accuracy = 0.773007
   Evaluation accuracy = 0.779948, Elapsed Time = 157.615884s
   Starting Epoch 6:
   Training loss = 910.041748, training accuracy = 0.796935
   Evaluation accuracy = 0.803886, Elapsed Time = 157.282300s
   Starting Epoch 7:
   Training loss = 829.032227, training accuracy = 0.817882
   Evaluation accuracy = 0.810296, Elapsed Time = 157.427150s
   Starting Epoch 8:
   Training loss = 761.830078, training accuracy = 0.832686
   Evaluation accuracy = 0.825621, Elapsed Time = 157.431677s
   Starting Epoch 9:
   Training loss = 703.944153, training accuracy = 0.846791
   Evaluation accuracy = 0.825421, Elapsed Time = 157.413234s
   ```

----------------------------------------------------------------
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
 
For queries about this service, please contact Infrastructure at:
users@infra.apache.org


With regards,
Apache Git Services

Mime
View raw message