GPU Track Model and Documentation

[Abhishek Thakur]

Hi All,

EC2 AMI information:

AMI ID: ami-a025c4cd

AMI Name: automl_gpu_abhishek

AMI Region: Northern Virginia (US East, N.Virginia)

To run the code:

./run.sh  (from the home directory)

Submission, ready to be submitted to Codalab, will be available in home directory as submission.zip

Python code is located at: https://github.com/abhishekkrthakur/automl_gpu

Initial explaination can be found here: https://www.linkedin.com/pulse/keras-neural-networks-win-nvidia-titan-x-abhishek-thakur

The datasets are already included as pickle files in the ec2 image. 

To dump datasets again, please uncomment the following line in ~/run.sh:

# Uncomment line below if you dont have a dataset dump as pickles

#python dump_datasets.py

The whole thing takes a little less than 1hr 40mins to run (including dumping of datasets).

In case of any problems, please let me know.

Thank you,

Regards,

Abhishek