awesome-deep-learning-finetune-experience

This repository not only contains experience about parameter finetune, but also other in-practice experience such as model ensemble (boosting, bagging and stacking) in Kaggle or other competitions.

Welcome pull request!

General

• My Neural Network isn't working! What should I do?
• williamFalcon/DeepRLHacks: Hacks for training RL systems from John Schulman's lecture at Deep RL Bootcamp (Aug 2017)
• 浅谈深度学习中的代码效率问题

Course

• How to Win a Data Science Competition: Learn from Top Kagglers | Coursera

Stacking & Boosting & Ensemble

• 模型融合方法概述 | 知乎专栏
• 如何在 Kaggle 首战中进入前 10% | Wille
• Introduction to Ensembling/Stacking in Python | Kaggle
• 一文读懂集成学习（附学习资源） | 数据派THU

Hyper-parameter search

• Scikit-learn hyperparameter search wrapper | skopt API documentation

GAN

• 生成对抗网络综述：从架构到训练技巧，看这篇论文就够了 | 机器之心 [英文原文]

Computer Vision

Image Classification

• [CVPR 2017] Squeeze-and-Excitation networks (ILSVRC 2017 winner) at CVPR2017  
• 专栏 | Momenta详解ImageNet 2017夺冠架构SENet
• [FAIR 2017] Accurate, Large Minibatch SGD: Training ImageNet in 1 Hour
• Kaggle Invasive Species Monitor competition : 3rd place solution overview

multi-label classification

• Kaggle亚马逊比赛冠军专访：利用标签相关性来处理分类问题 | 雷锋网

Detection

• 天池医疗AI大赛 第一季 宜远智能_HKBU团队Rank7 解决方案（含代码） [code]

Movie Description

• [LSMDC 2016] 1st place: Video Description by Combining Strong Representation and a Simple Nearest Neighbor Approach. Gil Levi, Dotan Kaufman, Lior Wolf (Tel Aviv University), Tal Hassner (University of Southern California) [slides]

Movie Annotation and Retrieval

• [LSMDC 2016] 1st place: Video Captioning and Retrieval Models with Semantic Attention. YoungJae Yu, Hyungjin Ko, Jongwook Choi, Gunhee Kim (Seoul National University) [slides]
• [LSMDC 2016] 2st place: Video Description by Combining Strong Representation and a Simple Nearest Neighbor Approach. Gil Levi, Dotan Kaufman, Lior Wolf (Tel Aviv University), Tal Hassner (University of Southern California)

Movie Fill-in-the-Blank

• [LSMDC 2016] 1st place: Video Captioning and Retrieval Models with Semantic Attention. YoungJae Yu, Hyungjin Ko, Jongwook Choi, Gunhee Kim (Seoul National University)

• [LSMDC 2016] 2st place: Video Fill in the Blank with Merging LSTMs. Amir Mazaheri, Dong Zhang, Mubarak Shah (University of Central Florida)

Speech

• 10686 一次 CTC-RNN 调参经历 | 知乎专栏

Feature Engineering

• 千锤万凿出深山：且谈特征工程最佳实践 | 大数据杂谈

• plantsgo/Rental-Listing-Inquiries: kaggle:Two Sigma Connect: Rental Listing Inquiries--top1