/Multi-label-Image-Classification-using-Tensorflow

Implementation of: 1) simple CNN for MNIST 2) Alexnet and VGG16 net(from scratch as well as using pre-trained ImageNet weights) on Pascal VOC2007 dataset

Primary LanguagePython

Multi-label Image Classification using Tensorflow

Implementation of simple CNN on MNIST, VGG16 and Alexnet on Pascal VOC dataset

  1. 00_mnist.py: Contains code for MNIST 10-digit classification in Tensorflow

  2. 01_pascal.py: CNN architecture for MNIST on Pascal VOC dataset

  3. 02_pascal_alexnet.py: Alexnet on Pascal VOC

  4. 03_pascal_vgg16.py: VGG16 on Pascal VOC from scratch

  5. 04_pascal_vggfinetune.py : Fine-tuning VGG16 on Pascal VOC using pre-trained weights

  6. 5a_conv1.py: Script to generate conv1 visualisation features. gist_cifar10_train.py : Needed to run 5a_conv1.py

i) Place 5a_conv1.py in the created directory containing the ckpt files(obtained from train). ii) Run 5a_conv1.py to obtain a folder containing the tensor board object. iii) Run tensor board —logdir= image_filters