An implementation of DeepMind's Relational Recurrent Neural Networks in PyTorch.

Overview

relational-rnn-pytorch

An implementation of DeepMind's Relational Recurrent Neural Networks (Santoro et al. 2018) in PyTorch.

Relational Memory Core (RMC) module is originally from official Sonnet implementation. However, currently they do not provide a full language modeling benchmark code.

This repo is a port of RMC with additional comments. It features a full-fledged word language modeling benchmark vs. traditional LSTM.

It supports any arbitrary word token-based text dataset, including WikiText-2 & WikiText-103.

Both RMC & LSTM models support adaptive softmax for much lower memory usage of large vocabulary dataset. RMC supports PyTorch's DataParallel, so you can easily experiment with a multi-GPU setup.

benchmark codes are hard-forked from official PyTorch word-language-model example

It also features an N-th farthest synthetic task from the paper (see below).

Requirements

PyTorch 0.4.1 or later (Tested on 1.0.0) & Python 3.6

Examples

python train_rmc.py --cuda for full training & test run of RMC with GPU.

python train_rmc.py --cuda --adaptivesoftmax --cutoffs 1000 5000 20000 if using large vocabulary dataset (like WikiText-103) to fit all the tensors in the VRAM.

python generate_rmc.py --cuda for generating sentences from the trained model.

python train_rnn.py --cuda for full training & test run of traditional RNN with GPU.

All default hyperparameters of RMC & LSTM are results from a two-week experiment using WikiText-2.

Data Preparation

Tested with WikiText-2 and WikiText-103. WikiText-2 is bundled.

Create a subfolder inside ./data and place word-level train.txt, valid.txt, and test.txt inside the subfolder.

Specify --data=(subfolder name) and you are good to go.

The code performs tokenization at the first training run, and the corpus is saved as pickle. The code will load the pickle file after the first run.

WikiText-2 Benchmark Results

Both RMC & LSTM have ~11M parameters. Please refer to the training code for details on hyperparameters.

Models Valid Perplexity Test Perplexity Forward pass ms/batch (TITAN Xp) Forward pass ms/batch (TITAN V)
LSTM (CuDNN) 111.31 105.56 26~27 40~41
LSTM (For Loop) Same as CuDNN Same as CuDNN 30~31 60~61
RMC 112.77 107.21 110~130 220~230

RMC can reach a comparable performance to LSTM (with heavy hyperparameter search), but it turns out that the RMC is very slow. The multi-head self-attention at every time step may be the culprit here. Using LSTMCell with for loop (which is more "fair" benchmark for RMC) slows down the forward pass, but it's still much faster.

Please also note that the hyperparameter for RMC is a worst-case scenario in terms of speed, because it used a single memory slot (as described in the paper) and did not benefit from a row-wise weight sharing from multi-slot memory.

Interesting to note here is that the speed is slower in TITAN V than TITAN Xp. The reason might be that the models are relatively small and the model calls small linear operations frequently.

Maybe TITAN Xp (~1,900Mhz unlocked CUDA clock speed vs. TITAN V's 1,335Mhz limit) benefits from these kind of workload. Or maybe TITAN V's CUDA kernel launch latency is higher for the ops in the model.

I'm not an expert in details of CUDA. Please share your results!

RMC Hyperparameter Search Results

Attention parameters tend to overfit the WikiText-2. reducing the hyperparmeters for attention (key_size) can combat the overfitting.

Applying dropout at the output logit before the softmax (like the LSTM one) helped preventing the overfitting.

embed & head size # heads attention MLP layers key size dropout at output memory slots test ppl
128 4 3 128 No 1 128.81
128 4 3 128 No 1 128.81
128 8 3 128 No 1 141.84
128 4 3 32 No 1 123.26
128 4 3 32 Yes 1 112.4
128 4 3 64 No 1 124.44
128 4 3 64 Yes 1 110.16
128 4 2 64 Yes 1 111.67
64 4 3 64 Yes 1 133.68
64 4 3 32 Yes 1 135.93
64 4 3 64 Yes 4 137.93
192 4 3 64 Yes 1 107.21
192 4 3 64 Yes 4 114.85
256 4 3 256 No 1 194.73
256 4 3 64 Yes 1 126.39

About WikiText-103

The original RMC paper presents WikiText-103 results with a larger model & batch size (6 Tesla P100, each with 64 batch size, so a total of 384. Ouch).

Using a full softmax easily blows up the VRAM. Using --adaptivesoftmax is highly recommended. If using --adaptivesoftmax, --cutoffs should be properly provided. Please refer to the original API description

I don't have such hardware and my resource is too limited to do the experiments. Benchmark result, or any other contributions are very welcome!

Nth Farthest Task

The objective of the task is: Given k randomly labelled (from 1 to k) D-dimensional vectors, identify which is the Nth farthest vector from vector M. (The answer is an integer from 1 to k.)

The specific task in the paper is: given 8 labelled 16-dimensional vectors, which is the Nth farthest vector from vector M? The vectors are labelled randomly so the model has to recognise that the Mth vector is the vector labelled as M as opposed to the vector in the Mth position in the input.

The input to the model comprises 8 40-dimensional vectors for each example. Each of these 40-dimensional vectors is structured like this:

[(vector 1) (label: which vector is it, from 1 to 8, one-hot encoded) (N, one-hot encoded) (M, one-hot encoded)] 

Example

python train_nth_farthest.py --cuda for training and testing on the Nth Farthest Task with GPU(s).

This uses the RelationalMemory class in relational_rnn_general.py, which is a version of relational_rnn_models.py without the language-modelling specific code.

Please refer totrain_nth_farthest.py for details on hyperparameter values. These are taken from Appendix A1 in the paper and from the Sonnet implementation when the hyperparameter values are not given in the paper.

Note: new examples are generated per epoch as in the Sonnet implementation. This seems to be consistent with the paper, which does not specify the number of examples used.

Experiment results

The model has been trained with a single TITAN Xp GPU for forever until it reaches 91% test accuracy. Below are the results with 3 independent runs:

The model does break the 25% barrier if trained long enough, but the wall clock time is roughly over 2~3x longer than those reported in the paper.

TODO

Experiment with different hyperparameters

Owner
Sang-gil Lee
Ph.D. student in ML/AI @ Seoul National University, South Korea. I do deep learning for sequence & generative models.
Sang-gil Lee
L-Verse: Bidirectional Generation Between Image and Text

Far beyond learning long-range interactions of natural language, transformers are becoming the de-facto standard for many vision tasks with their power and scalabilty

Kim, Taehoon 102 Dec 21, 2022
Code for database and frontend of webpage for Neural Fields in Visual Computing and Beyond.

Neural Fields in Visual Computing—Complementary Webpage This is based on the amazing MiniConf project from Hendrik Strobelt and Sasha Rush—thank you!

Brown University Visual Computing Group 29 Nov 30, 2022
LinkNet - This repository contains our Torch7 implementation of the network developed by us at e-Lab.

LinkNet This repository contains our Torch7 implementation of the network developed by us at e-Lab. You can go to our blogpost or read the article Lin

e-Lab 158 Nov 11, 2022
Approaches to modeling terrain and maps in python

topography 🌎 Contains different approaches to modeling terrain and topographic-style maps in python Features Inverse Distance Weighting (IDW) A given

John Gutierrez 1 Aug 10, 2022
Efficient Training of Visual Transformers with Small Datasets

Official codes for "Efficient Training of Visual Transformers with Small Datasets", NerIPS 2021.

Yahui Liu 112 Dec 25, 2022
​ This is the Pytorch implementation of Progressive Attentional Manifold Alignment.

PAMA This is the Pytorch implementation of Progressive Attentional Manifold Alignment. Requirements python 3.6 pytorch 1.2.0+ PIL, numpy, matplotlib C

98 Nov 15, 2022
The MATH Dataset

Measuring Mathematical Problem Solving With the MATH Dataset This is the repository for Measuring Mathematical Problem Solving With the MATH Dataset b

Dan Hendrycks 267 Dec 26, 2022
Tensorflow implementation of Semi-supervised Sequence Learning (https://arxiv.org/abs/1511.01432)

Transfer Learning for Text Classification with Tensorflow Tensorflow implementation of Semi-supervised Sequence Learning(https://arxiv.org/abs/1511.01

DONGJUN LEE 82 Oct 22, 2022
A neuroanatomy-based augmented reality experience powered by computer vision. Features 3D visuals of the Atlas Brain Map slices.

Brain Augmented Reality (AR) A neuroanatomy-based augmented reality experience powered by computer vision that features 3D visuals of the Atlas Brain

Yasmeen Brain 10 Oct 06, 2022
This is a Keras implementation of a CNN for estimating age, gender and mask from a camera.

face-detector-age-gender This is a Keras implementation of a CNN for estimating age, gender and mask from a camera. Before run face detector app, expr

Devdreamsolution 2 Dec 04, 2021
A repository for benchmarking neural vocoders by their quality and speed.

License The majority of VocBench is licensed under CC-BY-NC, however portions of the project are available under separate license terms: Wavenet, Para

Meta Research 177 Dec 12, 2022
U-Net: Convolutional Networks for Biomedical Image Segmentation

Deep Learning Tutorial for Kaggle Ultrasound Nerve Segmentation competition, using Keras This tutorial shows how to use Keras library to build deep ne

Yihui He 401 Nov 21, 2022
Machine Learning Privacy Meter: A tool to quantify the privacy risks of machine learning models with respect to inference attacks, notably membership inference attacks

ML Privacy Meter Machine learning is playing a central role in automated decision making in a wide range of organization and service providers. The da

Data Privacy and Trustworthy Machine Learning Research Lab 357 Jan 06, 2023
The trained model and denoising example for paper : Cardiopulmonary Auscultation Enhancement with a Two-Stage Noise Cancellation Approach

The trained model and denoising example for paper : Cardiopulmonary Auscultation Enhancement with a Two-Stage Noise Cancellation Approach

ycj_project 1 Jan 18, 2022
Rocket-recycling with Reinforcement Learning

Rocket-recycling with Reinforcement Learning Developed by: Zhengxia Zou I have long been fascinated by the recovery process of SpaceX rockets. In this

Zhengxia Zou 202 Jan 03, 2023
Deep Learning and Reinforcement Learning Library for Scientists and Engineers 🔥

TensorLayer is a novel TensorFlow-based deep learning and reinforcement learning library designed for researchers and engineers. It provides an extens

TensorLayer Community 7.1k Dec 29, 2022
The project page of paper: Architecture disentanglement for deep neural networks [ICCV 2021, oral]

This is the project page for the paper: Architecture Disentanglement for Deep Neural Networks, Jie Hu, Liujuan Cao, Tong Tong, Ye Qixiang, ShengChuan

Jie Hu 15 Aug 30, 2022
Gans-in-action - Companion repository to GANs in Action: Deep learning with Generative Adversarial Networks

GANs in Action by Jakub Langr and Vladimir Bok List of available code: Chapter 2: Colab, Notebook Chapter 3: Notebook Chapter 4: Notebook Chapter 6: C

GANs in Action 914 Dec 21, 2022
This is the official repository of XVFI (eXtreme Video Frame Interpolation)

XVFI This is the official repository of XVFI (eXtreme Video Frame Interpolation), https://arxiv.org/abs/2103.16206 Last Update: 20210607 We provide th

Jihyong Oh 195 Dec 29, 2022
Official implementation of "CrossPoint: Self-Supervised Cross-Modal Contrastive Learning for 3D Point Cloud Understanding" (CVPR, 2022)

CrossPoint: Self-Supervised Cross-Modal Contrastive Learning for 3D Point Cloud Understanding (CVPR'22) Paper Link | Project Page Abstract : Manual an

Mohamed Afham 152 Dec 23, 2022