Dog breed identification with gluon

Introduction

Dog Breed Identification is a kaggle playground prediction competition for determining the breed of a dog in an image.

Just use the fining-tuning models for features extraction, then build a very sample full-connections layer to classify. It get the the 69th (top5)

Clone this repository git clone git@github.com:classtag/dlnd_dog_breed_identification.git

Create conda envoriment

Please prepare your annacoda envoriment by Google.

conda create -n dlnd_dog_breed_identification python=3

Install packages

pip install -U --pre mxnet-cu80
pip install -U --pre jupyter

Open this repository files with your jupyter.

Name		Name	Last commit message	Last commit date
Latest commit History 7 Commits
.gitignore		.gitignore
1_dataset_prepare.ipynb		1_dataset_prepare.ipynb
1_kaggle_dataset_prepare_kaggle.ipynb		1_kaggle_dataset_prepare_kaggle.ipynb
2_extract_features.ipynb		2_extract_features.ipynb
2_extract_features.py		2_extract_features.py
2_extract_features_deprecate.ipynb		2_extract_features_deprecate.ipynb
2_feature_extractor_with_pretrain_models.py		2_feature_extractor_with_pretrain_models.py
2_model_selection.py		2_model_selection.py
3_build_classifier.ipynb		3_build_classifier.ipynb
3_classifier.ipynb		3_classifier.ipynb
Model Flow Diagram.png		Model Flow Diagram.png
README.md		README.md
rank.png		rank.png
stanford_lables.csv		stanford_lables.csv