Code to reproduce "imagenet in 18 minutes" experiments
Switch branches/tags
Nothing to show
Clone or download
Latest commit 6494420 Oct 16, 2018
Permalink
Failed to load latest commit information.
dawn Revert "Pep8" Oct 17, 2018
tools Convert async --> non_blocking for Python 3.7 Sep 11, 2018
training Revert "Pep8" Oct 17, 2018
.gitignore Adding single machine logs Sep 12, 2018
LICENSE add license Oct 4, 2018
README.md Update README.md Oct 8, 2018
requirements.txt First commit Sep 7, 2018
setup.sh First commit Sep 7, 2018
tensorboard.png . Sep 8, 2018
train.py Revert "Pep8" Oct 17, 2018

README.md

Code to reproduce ImageNet in 18 minutes, by Andrew Shaw, Yaroslav Bulatov, and Jeremy Howard. High-level overview of techniques used is here

Pre-requisites: Python 3.6 or higher

pip install -r requirements.txt
aws configure  (or set your AWS_ACCESS_KEY_ID/AWS_SECRET_ACCESS_KEY/AWS_DEFAULT_REGION)
python train.py  # pre-warming
python train.py 

To run with smaller number of machines:

python train.py --machines=1
python train.py --machines=4
python train.py --machines=8
python train.py --machines=16

Your AWS account needs to have high enough limit in order to reserve this number of p3.16xlarge instances. The code will set up necessary infrastructure like EFS, VPC, subnets, keypairs and placement groups. Therefore permissions for these those resources are needed.

Checking progress

Machines print progress to local stdout as well as logging TensorBoard event files to EFS. You can:

  1. launch tensorboard using tools/launch_tensorboard.py

That will provide a link to tensorboard instance which has loss graph under "losses" group. You'll see something like this under "Losses" tab

  1. Connect to one of the instances using instructions printed during launch. Look for something like this
2018-09-06 17:26:23.562096 15.imagenet: To connect to 15.imagenet
ssh -i /Users/yaroslav/.ncluster/ncluster5-yaroslav-316880547378-us-east-1.pem -o StrictHostKeyChecking=no ubuntu@18.206.193.26
tmux a

This will connect you to tmux session and you will see something like this

.997 (65.102)   Acc@5 85.854 (85.224)   Data 0.004 (0.035)      BW 2.444 2.445
Epoch: [21][175/179]    Time 0.318 (0.368)      Loss 1.4276 (1.4767)    Acc@1 66.169 (65.132)   Acc@5 86.063 (85.244)   Data 0.004 (0.035)      BW 2.464 2.466
Changing LR from 0.4012569832402235 to 0.40000000000000013
Epoch: [21][179/179]    Time 0.336 (0.367)      Loss 1.4457 (1.4761)    Acc@1 65.473 (65.152)   Acc@5 86.061 (85.252)   Data 0.004 (0.034)      BW 2.393 2.397
Test:  [21][5/7]        Time 0.106 (0.563)      Loss 1.3254 (1.3187)    Acc@1 67.508 (67.693)   Acc@5 88.644 (88.315)
Test:  [21][7/7]        Time 0.105 (0.432)      Loss 1.4089 (1.3346)    Acc@1 67.134 (67.462)   Acc@5 87.257 (88.124)
~~21    0.31132         67.462          88.124

The last number indicates that at epoch 21 the run got 67.462 top-1 test accuracy and 88.124 top-5 test accuracy.