Loading presentation...

Present Remotely

Send the link below via email or IM

Copy

Present to your audience

Start remote presentation

  • Invited audience members will follow you as you navigate and present
  • People invited to a presentation do not need a Prezi account
  • This link expires 10 minutes after you close the presentation
  • A maximum of 30 users can follow your presentation
  • Learn more about this feature in our knowledge base article

Do you really want to delete this prezi?

Neither you, nor the coeditors you shared it with will be able to recover it again.

DeleteCancel

Faster R-CNN

No description
by

Jon Rayner

on 16 September 2017

Comments (0)

Please log in to add your comment.

Report abuse

Transcript of Faster R-CNN

What?
- Combines a Region Proposal Network (RPN) and a Region-based Convolutional Network (R-CNN)
Zeros
- Wading through complicated distribution code with no time to train
Heros
- Running on AWS GPU instance, tried Azure as well
Chasing the Faster R-CNN
William Grunow, Sicelukwanda Zwane,
Zainoolabadien Karim, Jonathan Rayner
- SOTA accuracy and faster than previous R-CNN's
- In principle the code works, but more training time needed
- Gunzip, MTN, and Pizza
- "Running" 6 GB mem networks on 4 GB GPU's
- Uploading image data to the cloud, then deleting it accidentally, breaking TensorFlow several times, trying to run convnets on CPU's
Thanks for the competition
and to the organizers!
code used from https://github.com/yhenon/keras-frcnn
Full transcript