GitPlanet

Cheap and reliable Node.js hosting starts at $3/month, and $1/month static HTML hosting

Created with love in Canada, visit hostnodejs.com today

Feel like to post an Ad? Learn Details
All Projects → andy840314 → QANet-pytorch-

andy840314 / QANet-pytorch-

Licence: MIT license
A Pytorch implementation of QANet

Programming Languages

python
139335 projects - #7 most used programming language
shell
77523 projects

Labels

nlppytorchsquad

Projects that are alternatives of or similar to QANet-pytorch-

R Net
A Tensorflow Implementation of R-net: Machine reading comprehension with self matching networks
Stars: ✭ 321 (+252.75%)
Mutual labels:  squad
Bidaf Pytorch
An Implementation of Bidirectional Attention Flow
Stars: ✭ 42 (-53.85%)
Mutual labels:  squad
Albert Tf2.0
ALBERT model Pretraining and Fine Tuning using TF2.0
Stars: ✭ 180 (+97.8%)
Mutual labels:  squad
Lambdahack
Haskell game engine library for roguelike dungeon crawlers; please offer feedback, e.g., after trying out the sample game with the web frontend at
Stars: ✭ 439 (+382.42%)
Mutual labels:  squad
Reading comprehension tf
Machine Reading Comprehension in Tensorflow
Stars: ✭ 37 (-59.34%)
Mutual labels:  squad
Bi Att Flow
Bi-directional Attention Flow (BiDAF) network is a multi-stage hierarchical process that represents context at different levels of granularity and uses a bi-directional attention flow mechanism to achieve a query-aware context representation without early summarization.
Stars: ✭ 1,472 (+1517.58%)
Mutual labels:  squad
co-attention
Pytorch implementation of "Dynamic Coattention Networks For Question Answering"
Stars: ✭ 54 (-40.66%)
Mutual labels:  squad
R Net
R-net in PyTorch, with ELMo
Stars: ✭ 194 (+113.19%)
Mutual labels:  squad
Qanet
A Tensorflow implementation of QANet for machine reading comprehension
Stars: ✭ 996 (+994.51%)
Mutual labels:  squad
Allure
Allure of the Stars is a near-future Sci-Fi roguelike and tactical squad combat game written in Haskell; please offer feedback, e.g., after trying out the web frontend version at
Stars: ✭ 149 (+63.74%)
Mutual labels:  squad
R Net
Tensorflow Implementation of R-Net
Stars: ✭ 582 (+539.56%)
Mutual labels:  squad
Fusionnet
My implementation of the FusionNet for machine comprehension
Stars: ✭ 29 (-68.13%)
Mutual labels:  squad
Haystack
🔍 Haystack is an open source NLP framework that leverages Transformer models. It enables developers to implement production-ready neural search, question answering, semantic document search and summarization for a wide range of applications.
Stars: ✭ 3,409 (+3646.15%)
Mutual labels:  squad
Drqa
A pytorch implementation of Reading Wikipedia to Answer Open-Domain Questions.
Stars: ✭ 378 (+315.38%)
Mutual labels:  squad
R Net In Keras
Open R-NET implementation and detailed analysis: https://git.io/vd8dx
Stars: ✭ 181 (+98.9%)
Mutual labels:  squad
Learning to retrieve reasoning paths
The official implementation of ICLR 2020, "Learning to Retrieve Reasoning Paths over Wikipedia Graph for Question Answering".
Stars: ✭ 318 (+249.45%)
Mutual labels:  squad
Match Lstm
A PyTorch implemention of Match-LSTM, R-NET and M-Reader for Machine Reading Comprehension
Stars: ✭ 92 (+1.1%)
Mutual labels:  squad
Dawn Bench Entries
DAWNBench: An End-to-End Deep Learning Benchmark and Competition
Stars: ✭ 254 (+179.12%)
Mutual labels:  squad
X3daudio1 7 hrtf
HRTF for Arma 3, Skyrim, and other titles that use XAudio2 + X3DAudio
Stars: ✭ 192 (+110.99%)
Mutual labels:  squad
Mnemonicreader
A PyTorch implementation of Mnemonic Reader for the Machine Comprehension task
Stars: ✭ 137 (+50.55%)
Mutual labels:  squad
View All Similar Projects ➔

QANet

A Pytorch implementation of QANet
The code is mostly based on the two repositories: hengruo/QANet-pytorch NLPLearn/QANet

Performance

Training epochs / Steps BatchSize HiddenSize Attention Heads EM F1
12.8 / 35,000 32 96 1 69.0 78.6
22 / 60,000 32 96 1 69.7 79.2
12.8 / 93,200 12 128 8 70.3 79.7
22 / 160,160 12 128 8 70.7 80.0

*The results of hidden size 128 with 8 heads were run with 12 batches.

Requirements

  • python 3.6
  • pytorch 0.4.0
  • tqdm
  • spacy 2.0.11
  • tensorboardX
  • absl-py

Usage

Download and preprocess the data

# download SQuAD and Glove
$ sh download.sh
# preprocess
$ python3.6 main.py --mode data

Train the model

# model/model.pt will be generated every epoch
$ python3.6 main.py --mode train

Tensorboard

# Run tensorboard for visualisation
$ tensorboard --logdir ./log/

TODO

  • Add Exponential Moving Average
  • Reach the performance of the paper with hidden size 96, 1 head.
  • Test on hidden size 128, 8 head.
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].