Skip to main content

Uploading Large Files

Some models may require uploading large files into the containerization environment. To do this, we provide access to Amazon Web Service's Simple Cloud Storage (AWS S3). Accessing the Dojo S3 is easy, but it does require an AWS access key and secret key. Please reach out to to request access.

There are numerous ways to upload files to S3. We will describe below the two most common methods:

  1. Using the AWS CLI
  2. Using the Python boto3 library

WARNING: leaking your Dojo AWS credentials to the public may cause your access to Dojo to be revoked!

Uploading files with the AWS CLI

First, you'll need to download and install the AWS CLI. Next, in a terminal, type aws configure. You should be prompted for the AWS access key and secret key provided to you by the Dojo team.

If you are already using the AWS CLI you should consider adding a named profile called dojo-modeling.

Now, to upload a file you can run:

aws s3 cp my_large_file.csv s3://dojo-modeling-data/

If you have created an AWS profile locally you may append --profile=dojo-modeling to the above command. Once you've received confirmation that the file is uploaded, it will be available to the public at:

All files uploaded to the dojo-modeling-data S3 bucket are public by default.

Uploading files with boto3

If you are a Python user, you may wish to upload large files using Python's boto3 library. First install it with pip3 install boto3. Then, in a Python terminal run:

import boto3

s3_client = boto3.client(


You should now be able to download the file from

Downloading your large file

You may download your large file from the containerization environment using wget: