Progress for week 5 (2016)
From Robin
(Difference between revisions)
(→Eirik Sundet) |
(→Kai Wig) |
||
| Line 25: | Line 25: | ||
* Compare test accuracy against training iterations | * Compare test accuracy against training iterations | ||
* Plot ROC curves for the test data | * Plot ROC curves for the test data | ||
| - | |||
* Surgically remove neurons | * Surgically remove neurons | ||
| Line 32: | Line 31: | ||
* Done for reinitialized model and fully trained model up to iter #240K | * Done for reinitialized model and fully trained model up to iter #240K | ||
* Done | * Done | ||
| - | |||
* Delayed | * Delayed | ||
| + | |||
Extra: | Extra: | ||
| - | * | + | * Got the deep visualization toolbox up and running, next is the evolutionary add-on |
== [[user:Fredrsae|Fredrik]] == | == [[user:Fredrsae|Fredrik]] == | ||
Revision as of 14:28, 11 February 2016
Contents |
Student template (copy this for your entry)
Budget
- Todo 1
- Todo 2
Accounting
- Done 1
- Done 2
Eirik Sundet
Budget
- Thesis writing
Accounting
- Done 1
Also:
- Exported result images from C++/OpenCV to MATLAB
- Evaluating results: ROC Curves
- Explored other statistical measures
Kai Wig
Budget
- Further fine tune Kitti CNN model
- Compare test accuracy against training iterations
- Plot ROC curves for the test data
- Surgically remove neurons
Accounting
- Trained with/without locked Conv layers, with/without reinitialized top layers
- Done for reinitialized model and fully trained model up to iter #240K
- Done
- Delayed
Extra:
- Got the deep visualization toolbox up and running, next is the evolutionary add-on
Fredrik
Budget
- Implement IT&E
- Fix multithreading (low prio)
- Thesis writing (Things not related to results)
