New Amazon EC2 DL1 instances powered by Gaudi accelerators from Habana Labs, an Intel company, are designed specifically for training deep learning models. We invite you to test your skills and train a deep learning model of your choice using DL1 instances and Habana’s SynapseAI SDK. DL1 instances are powered by 8 Gaudi accelerators and we encourage you to utilize multiple Gaudi accelerators to efficiently scale your deep learning model training.

 

What to Build

Train a deep learning model on new Amazon EC2 DL1 instances. You can also migrate existing models to DL1 instances. 

What to Submit

  • Provide a URL to your code repository for judging and testing. Provide access by making it public or share access as instructed in the submission form.
  • Explain and identify the datasets used in training your model.
  • Include a video (about 3 minutes) that demonstrates your submission. Videos must be uploaded to YouTube, Vimeo, or Facebook Video and made public.
  • Provide email associated with the AWS Account ID.

Hackathon Sponsors

Prizes

$150,000 in prizes

First Place

• $35,000 USD
• $5,000 in AWS Credits

Second Place

• $20,000 USD
• $5,000 in AWS Credits

Third Place

• $15,000 USD
• $5,000 in AWS Credits

Fourth Place

• $10,000 USD
• $5,000 in AWS Credits

Best Computer Vision Application

• $5,000 USD
• $1,000 in AWS Credits

Runner up Computer Vision Applications (3)

• $2000 USD
• $500 in AWS Credits

Best Natural Language Processing Applications

• $5,000 USD
• $1,000 in AWS Credits

Runner Ups Natural Language Processing Applications (3)

• $2000 USD
• $500 in AWS Credits

Best Technical Implementation

• $5,000 USD
• $1,000 in AWS Credits

Highest Potential Impact

• $5,000 USD
• $1,000 in AWS Credits

Most Creative

• $5,000 USD
• $1,000 in AWS Credits

Eligible Project Submitter (100)

Credits for AWS merchandise worth up to $50 for the first 100 eligible project team members.

Devpost Achievements

Submitting to this hackathon could earn you:

Judges

To be announced...

To be announced...

Judging Criteria

  • 20% - Quality of the Idea
    Includes creativity, originality, and code quality.
  • 40% - Potential Value
    Best practices and techniques that can help the broader community use DL1.
  • 40% - Implementation
    How well is the idea executed by the developers and the efficiency of their code in utilizing the Gaudi accelerators? Extra points if your model is trained on more than one accelerator.

Questions? Email the hackathon manager

Invite others to compete

This site is protected by reCAPTCHA and the Google Privacy Policy and Terms of Service apply.