Training Xception model for Kaggle competition “Cdiscount’s Image Classification Challenge”
نویسنده
چکیده
This commutation is about training the Xception model for the Kaggle competition “Cdiscount’s Image Classification Challenge”. The paper will briefly describe all methods/code (github.com/ardiloot/CDiscountClassifier) used to train the model for best classification performance. Mainly, the effect of the augmentation (both train and test time) and algebraic ensemble methods were studied. In the competition, the model without augmentation, no test-timeaugmentation and with power law ensemble method scored the best (accuracy 0.72582) and guaranteed the 64th position out of 627 teams.
منابع مشابه
Crime Prediction in San Fransisco
In June 2015, Kaggle began a competition named “San Francisco Crime Classification”[8], ending in June 2016. The competition’s dataset caught our attention due the subject being very tangible, with crime being at the forefront of modern media and to San Francisco being culturally significant due to its current tech industry. The dataset is also described by geographic and temporal features, the...
متن کاملThe Prediction of Booking Destination On Airbnb Dataset
This report is about analysis of the Airbnb dataset and the model we built to do the prediction task on the dataset. The dataset comes from an ongoing kaggle competition supported by Airbnb. We first did some comprehensive analysis on the dataset, explored most features and collected all features we thought was useful. Then we described and interpreted the prediction task and the evaluation met...
متن کاملSatellite Imagery Feature Detection using Deep Convolutional Neural Network: A Kaggle Competition
This paper describes our approach to the DSTL Satellite Imagery Feature Detection challenge [11] run by Kaggle. The primary goal of this challenge is accurate semantic segmentation of different classes in satellite imagery. Our approach is based on an adaptation of fully convolutional neural network for multispectral data processing. In addition, we defined several modifications to the training...
متن کاملPredicting Galaxy Morphology Classifications
I implement several DSP technique: FFT, Wiener Filter, convolution to preprocess the galaxy images and use some machine learning theories to learn and predict hand-made galaxy morphology classifications. My data and success metric are taken from a recent Kaggle competition “Galaxy Zoo The Galaxy Challenge”. I focus on three sequential algorithm modules: image analysis, data compression, and mac...
متن کاملThe YouTube-8M Kaggle Competition: Challenges and Methods
We took part in the YouTube-8M Video Understanding Challenge hosted on Kaggle, and achieved the 10th place within less than one month’s time. In this paper, we present an extensive analysis and solution to the underlying machine-learning problem based on frame-level data, where major challenges are identified and corresponding preliminary methods are proposed. It’s noteworthy that, with merely ...
متن کامل