All Projects → Ai-Light → Kdd2020multimodalities

Ai-Light / Kdd2020multimodalities

KDD Cup 2020 Challenges for Modern E-Commerce Platform: Multimodalities Recall

Projects that are alternatives of or similar to Kdd2020multimodalities

Deej A.i.
Create automatic playlists by using Deep Learning to *listen* to the music
Stars: ✭ 57 (-5%)
Mutual labels:  jupyter-notebook
Home price estimator
Uses Zillow metadata, NLP on realtor description, and VGG16 on home images to predict home sale prices in Portland from 6/16 - 7/17.
Stars: ✭ 59 (-1.67%)
Mutual labels:  jupyter-notebook
Intro To Nltk
Code and Notebooks for the Natural Language Processing with Python course.
Stars: ✭ 59 (-1.67%)
Mutual labels:  jupyter-notebook
Road Lane Instance Segmentation Pytorch
tuSimple dataset road lane instance segmentation with PyTorch, ROS, ENet, SegNet and Discriminative Loss.
Stars: ✭ 59 (-1.67%)
Mutual labels:  jupyter-notebook
Pydata 2018
Stars: ✭ 59 (-1.67%)
Mutual labels:  jupyter-notebook
Deep Learning 101
The tools and syntax you need to code neural networks from day one.
Stars: ✭ 59 (-1.67%)
Mutual labels:  jupyter-notebook
College Majors Visualisation
Interactive Data Visualisation of U.S College Majors
Stars: ✭ 59 (-1.67%)
Mutual labels:  jupyter-notebook
Image Outpainting
🏖 Keras Implementation of Painting outside the box
Stars: ✭ 1,106 (+1743.33%)
Mutual labels:  jupyter-notebook
Applied Text Mining In Python
Repo for Applied Text Mining in Python (coursera) by University of Michigan
Stars: ✭ 59 (-1.67%)
Mutual labels:  jupyter-notebook
Stylist
Fast artistic style transfer with convolutional neural networks.
Stars: ✭ 59 (-1.67%)
Mutual labels:  jupyter-notebook
Content based movie recommender
Stars: ✭ 59 (-1.67%)
Mutual labels:  jupyter-notebook
Rastermap
A multi-dimensional embedding algorithm
Stars: ✭ 58 (-3.33%)
Mutual labels:  jupyter-notebook
Learning Pandas Second Edition
Learning pandas, Second Edition, published by Packt
Stars: ✭ 59 (-1.67%)
Mutual labels:  jupyter-notebook
Ssd keras
Port of Single Shot MultiBox Detector to Keras
Stars: ✭ 1,101 (+1735%)
Mutual labels:  jupyter-notebook
Feedinlib
This repository contains implementations of photovoltaic models to calculate electricity generation from a pv installation based on given solar radiation. Furthermore it contains all necessary pre-calculations.
Stars: ✭ 59 (-1.67%)
Mutual labels:  jupyter-notebook
Pytorchdiscreteflows
Discrete Normalizing Flows implemented in PyTorch
Stars: ✭ 59 (-1.67%)
Mutual labels:  jupyter-notebook
Storytelling With Data
Course materials for Dartmouth Course: Storytelling with Data (PSYC 81.09).
Stars: ✭ 59 (-1.67%)
Mutual labels:  jupyter-notebook
Analytics vidhya
Codes related to activities on AV including articles, hackathons and discussions.
Stars: ✭ 1,106 (+1743.33%)
Mutual labels:  jupyter-notebook
Soccer xg
A Python package for training and analyzing expected goals (xG) models in soccer.
Stars: ✭ 60 (+0%)
Mutual labels:  jupyter-notebook
Keras model compression
Model Compression Based on Geoffery Hinton's Logit Regression Method in Keras applied to MNIST 16x compression over 0.95 percent accuracy.An Implementation of "Distilling the Knowledge in a Neural Network - Geoffery Hinton et. al"
Stars: ✭ 59 (-1.67%)
Mutual labels:  jupyter-notebook

KDD2020Multimodalities

KDD Cup 2020 Challenges for Modern E-Commerce Platform: Multimodalities Recall

Baseline Score: 0.5538(使用四分之一数据) | 0.599(使用一半的数据)

方案详解:https://zhuanlan.zhihu.com/p/135984016

https://fasttext.cc/docs/en/english-vectors.html crawl-300d-2M.vec.zip

https://nlp.stanford.edu/projects/glove/ glove.840B.300d.zip

https://github.com/google-research/bert BERT-Base, Uncased

数据全部存放在data目录下

运行顺序

1.preprocess.py(只读了1w的数据,要读取更多数据,谨慎修改函数pd.read_csv的参数,有大内存机器的忽略)

2.[image-concat-query]-wwm_uncased_L12-768_v3_quart.ipynb

环境和依赖

python 3.7

tensorflow==1.13.1

Keras==2.2.4

keras-bert==0.78.0

Note that the project description data, including the texts, logos, images, and/or trademarks, for each open source project belongs to its rightful owner. If you wish to add or remove any projects, please contact us at [email protected].