Difference between revisions of "DSC180F20W7"
From Immersive Visualization Lab Wiki
| Line 13: | Line 13: | ||
* [https://medium.com/howtoai/pytorch-torchvision-coco-dataset-b7f5e8cad82 Import the COCO dataset using the torchvision package.] | * [https://medium.com/howtoai/pytorch-torchvision-coco-dataset-b7f5e8cad82 Import the COCO dataset using the torchvision package.] | ||
* Teach a convolutional neural network image classification, using the COCO dataset. | * Teach a convolutional neural network image classification, using the COCO dataset. | ||
| − | * Implement a demo for your image classification algorithm | + | * Implement a demo for your image classification algorithm. |
| + | |||
| + | ==Report Portion== | ||
| + | |||
| + | * Describe the classification problem you set out to implement. | ||
| + | * Describe the approach of your solution. | ||
| + | * How accurate are the results of your image classification demo? (compare against ground truth) | ||
| + | * Suggest at least 3 things one could try to get more accurate results. (eg, network design, training data, training parameters, etc.) | ||
Revision as of 22:08, 11 November 2020
Replication Checkpoint #2
UNDER CONSTRUCTION
For Checkpoint 2 we are going to add deep learning to our toolkit, in order to eventually apply the Grad-CAM XAI techniques to the COCO dataset (but that will not happen until the final checkpoint).
PyTorch is a machine learning API which builds on top of Torch. Its most developed interface is in Python.
Code Portion
For this checkpoint you need to create a PyTorch application in Python to do the following things:
- Import the COCO dataset using the torchvision package.
- Teach a convolutional neural network image classification, using the COCO dataset.
- Implement a demo for your image classification algorithm.
Report Portion
- Describe the classification problem you set out to implement.
- Describe the approach of your solution.
- How accurate are the results of your image classification demo? (compare against ground truth)
- Suggest at least 3 things one could try to get more accurate results. (eg, network design, training data, training parameters, etc.)
