Skip to content
No description, website, or topics provided.
Python Shell
Branch: master
Clone or download
Fetching latest commit…
Cannot retrieve the latest commit at this time.
Permalink
Type Name Latest commit message Commit time
Failed to load latest commit information.
task1
task2
task3
task4
task5
utils
.gitignore
DLCV04_slides.pdf
README.md

README.md

dlcv04

This is the repository for the group 4 for the seminar of Deep Learning for Computer Vision in ETSETB:

https://github.com/imatge-upc/telecombcn-2016-dlcv/

And the group is formed by:

Manel Baradad, Míriam Bellver, Martí Cervià, Hector Esteban, Carlos Roig

The slides of the project can be seen here:

https://docs.google.com/presentation/d/1daS4M7e5Grk6Ytqk2kdapNonDVCPKwiv5HxscIA8UQI/edit?usp=sharing

We have performed 5 different tasks, described in the following lines:

TASK 1: ARCHITECTURE

  1. Build network for classification problem
  2. Study memory requirements and computational loads for different layers

Main idea: use MNIST and start with small network, test different layers and architectures

TASK 2: TRAINING

  1. Study impact in performance of DATA AUGMENTATION, batches size, batch normalization
  2. Training validation curves
  3. Overfitting

TASK 3: VISUALIZATION

  1. Visualize filter responses from own and also pretrained net
  2. t-SNE
  3. Off-the-shelf AlexNet

TASK 4: TRANSFER LEARNING

  1. Train network on CIFAR10 and fine-tune for Terrassa Buildings 900
  2. Off-the-shelf convnet

TASK 5: OPEN PROJECT

  1. We have played with neural style, changing the features that encode the style and see the results.

DATASETS:

We have worked with three different Datasets:

MNIST CIFAR10 Terrassa-building-900 (https://imatge.upc.edu/web/resources/terrassa-buildings-900)

You can’t perform that action at this time.