chrishkchris commented on a change in pull request #651: Simply example APIs
URL: https://github.com/apache/singa/pull/651#discussion_r403479208
##########
File path: examples/autograd/cifar10_multiprocess.py
##########
@@ -17,28 +17,16 @@
# under the License.
#
-from singa import opt
from resnet_cifar10 import *
import multiprocessing
-
-def data_partition(dataset_x, dataset_y, rank_in_global, world_size):
- data_per_rank = dataset_x.shape[0] // world_size
- idx_start = rank_in_global * data_per_rank
- idx_end = (rank_in_global + 1) * data_per_rank
- return dataset_x[idx_start:idx_end], dataset_y[idx_start:idx_end]
-
-
if __name__ == '__main__':
# Generate a NCCL ID to be used for collective communication
nccl_id = singa.NcclIdHolder()
- sgd = opt.SGD(lr=0.005, momentum=0.9, weight_decay=1e-5)
-
- gpu_per_node = 4
- max_epoch = 100
- batch_size = 32
+ # number of GPUs to be used
+ gpu_per_node = 2
Review comment:
I added command line argument
```
root@877a3759b148:~/dcsysh/singa/examples/autograd# python3
mnist_multiprocess.py 2
Starting Epoch 0:
Training loss = 1029.400879, training accuracy = 0.628155
Evaluation accuracy = 0.882712, Elapsed Time = 1.923681s
Starting Epoch 1:
Training loss = 367.118927, training accuracy = 0.874215
Evaluation accuracy = 0.939804, Elapsed Time = 1.986267s
Starting Epoch 2:
Training loss = 270.464111, training accuracy = 0.908504
Evaluation accuracy = 0.935497, Elapsed Time = 1.985472s
Starting Epoch 3:
Training loss = 215.307678, training accuracy = 0.927017
Evaluation accuracy = 0.958333, Elapsed Time = 1.894766s
Starting Epoch 4:
Training loss = 179.509125, training accuracy = 0.940371
Evaluation accuracy = 0.965244, Elapsed Time = 1.889032s
Starting Epoch 5:
Training loss = 157.717468, training accuracy = 0.947282
Evaluation accuracy = 0.962640, Elapsed Time = 1.892571s
Starting Epoch 6:
Training loss = 143.918320, training accuracy = 0.952224
Evaluation accuracy = 0.968950, Elapsed Time = 1.910212s
Starting Epoch 7:
Training loss = 134.035339, training accuracy = 0.954911
Evaluation accuracy = 0.957332, Elapsed Time = 1.977645s
Starting Epoch 8:
Training loss = 121.463905, training accuracy = 0.959352
Evaluation accuracy = 0.970252, Elapsed Time = 1.944209s
Starting Epoch 9:
Training loss = 112.466797, training accuracy = 0.962457
Evaluation accuracy = 0.974259, Elapsed Time = 1.883536s
```
----------------------------------------------------------------
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:
[email protected]
With regards,
Apache Git Services
