An implementation of the Pay Attention when Required transformer

Overview

Pay Attention when Required (PAR) Transformer-XL

An implementation of the Pay Attention when Required transformer from the paper: https://arxiv.org/pdf/2009.04534.pdf

alt text [source: Jonathan Kernes]

Quick overview

The Pay Attention when Required Transformer (Mandava, et. al. 2020) is just a regular transformer-XL (Dai et. al. 2019)[https://arxiv.org/pdf/1901.02860.pdf] , but the ratio of attention and dense layers has been optimized. This optimization is performed by allowing the network to choose which types of layer it prefers in each block of the network. The present implementation is not an exact replica of the author's efforts. Instead, we perform a simultaneous optimization procedure on both the model architecture and model parameters. The search is performed using a SuperNet, which is a sequential neural network composed of stochastic blocks, as shown in the figure below (taken from the paper. Please don't sue me!)

alt text [Source: Mandava et. al. 2020]

The key component is a Gumbel-Softmax layer [(Jang et al., 2016) and (Maddison et al., 2016). jang link: https://arxiv.org/pdf/1611.01144.pdf]. This layer is a continuous representation of a discrete sampling from a Categorical distribution, thereby allowing us to use gradients to learn parameters of a discrete distribution. (Recall a categorical is a distrbution over K states with kth state having probability pi_k, and we must have the normalization condition \sum_{i=1}^K pi_i = 1)

As the model learns, it is free to adjust both the usual model parameters, as well as its architecture search parameters pi, indicating the probability of choosing either

  1. Attention

  2. Dense

  3. Identity

for any given stochastic block. We perform simulated annealing: since the categorical distribution is approximated by a continuous representation, we get some scores like (0.02, 0.98, 0.02) for the probability of say sampling that state 2 is picked. The sharpness of this is set by a parameter \tau (the temperature), with a categorical distribution the limit tau-->0. Simulated annealing means we begin with tau=1 to let the model figure out what it wants, then slowly decrease tau so the distribution approaches a categorical.

All of this is implemented on the freely available wiki-text2 dataset.

Explanation of the main GIF: The main gif is the result of our experiments. It shows the pi distribution for each stochastic block of a 6 block SuperNet, as a function of training iterations. The number indicates the probability of the most likely layer type (darker means more probable). As you can see, the model learns to put attention in the beginning, and dense layers at the end.

Requirements

Usual ML stuff, if you have a conda environment, python 3+, TensorFlow 2+ you should be ok. You will need TensorFlow Text as well to handle the SentencePiece Tokenization

If you choose to run your own tokenizer (a flag option in data_utils for handling new text data), you will also need to download the SentencePiece package: https://github.com/google/sentencepiece

Data

The dataset used is Wiki-text2. We have provided a copy of this in the data folder, along with some preprocessed data for training. In order to reproduce this from scratch, run the shell script

./create_tfrecords.sh

This will download the wiki-text2 dataset from its source, then proceed to clean, batch, and write the data to a tfrecords file. The shell script calls build_data.py which offers more control over what type of data to generate. The general parameters you will want to tune are:

*batch_size *seq_len.

You can also supply your own dataset instead of the one provided. The underlying tokenizer uses sentencepiece (Kudo): https://github.com/google/sentencepiece, which works at the byte level and can handle any kind of input. Simply change the --input_text flag to your file, and set the desired --vocab_size.

Why do we need to specify the batch size? Transformer XL uses memory states to form a recurrent, long range network. After analyzing a particular sequence say [A,B] of the sequence [A,B,C,D], the results of [A,B] are fed into the [C,D] calculation with a stop gradient. Therefore, we must be sure that each datapoint follows chronologically from the previous one.

This is achieved by context batching (see data_utils.py function) where we break the entire dataset into batch_size segments, then pull in order one sequence from each batch at a time to form the dataset. Because of this, note that adding more shards to the data could result in a large loss (order of batch_size*seq_len*shards), as each shard will drop the remaining datapoint of size (batch_size*seq_len) to keep the tensor shapes.

Addtional technical details

Per the original Transformer-XL, we also implement an adaptive softmax layer (Grave et. al. 2017, https://arxiv.org/abs/1609.04309) to deal with a potentially large number of outputs in the final dense layer. This implemenation is inspired by the TF 1.0 example at https://github.com/yangsaiyong/tf-adaptive-softmax-lstm-lm. To use the adaptive softmax, set the --cutoffs= flag in train.py. The cutoffs are the max values of each bin, and should NOT include the vocab size (i.e. the max cutoff of the final bin). If no cutoffs are specified, the model defaults to normal softmax.

For completeness, we have also provided a script optimal_cuts.py that determines the optimal cutoffs given a return space separated file of unigram probabilities (based on the assumptions of Grave et. al. regarding GPU computation complexity -- see the paper for details). The algorithm uses dynamic programming, but is quite slow at O(KN^2), for K cutoffs and N vocab words. In principle it's a one time cost to determine the cutoffs, but we are impatient and recommend to just play around with the cutoffs instead. See the script for flag details

Training and Benchmarks

The default model we use has memory length 16, feed-forward dimension 1024, attention dimension 128, and 6 stochastic blocks, with an adaptive softmax layer and 2 clusters. We trained on a colab GPU for 20 epochs, taking a total of 37 minutes. We use an Adam optimzer with cosine rate decay: an initial warmup of 4000 steps and a maximum learning rate of 1e-4, decaying to zero at the end of training. Our training benchmarks are:

Iteration (thousands) Train_perplexity Validation_perplexity Time
2.7k 163.9 114.4 1m 58s
8.5k 78.56 62.33 5m 37s
14.1k 65.71 51.88 9m 28s
28.3k 48.52 42.61 18m 40s
48.1k 41.85 39.57 31m 51s
56.5k 42.12 39.41 37m 14s

To train, simply run the shell script

./base_model.sh

adjusting the parameters as you see fit. The above model is the default configuration. To train in colab, simply open up the notebook "colab.ipynb" and follow the instructions. This is most easily done by going to [google.colab.com] and searching this repository in github. The benefit of colab, is it's easier to play around with the model after training.

While training, we have provided two ways to monitor the output

  1. A tensorboard log. The colab notebook takes care of running this for you. In the terminal, first create a 'logs' directory, then run the command tensorboard --logdir logs in a separate tab. This will open a port where you can view live plots of the learning rate, tau annealing, train/valid loss and perplexity.

  2. An output log saved to training_log.log. This will log the model summary, parameters, etc. as well as print out loss updates every 100 steps and save it to the log file.

Thanks for reading this far!

Enjoy! And thank you to the wonderful researchers that inspired this project.

If you would like to contribute, or have any comments questions concerns please open a pull request or email me directly.

Demo programs for the Talking Head Anime from a Single Image 2: More Expressive project.

Demo Code for "Talking Head Anime from a Single Image 2: More Expressive" This repository contains demo programs for the Talking Head Anime

Pramook Khungurn 901 Jan 06, 2023
Pre-Training with Whole Word Masking for Chinese BERT

Pre-Training with Whole Word Masking for Chinese BERT

Yiming Cui 7.7k Dec 31, 2022
Summarization, translation, sentiment-analysis, text-generation and more at blazing speed using a T5 version implemented in ONNX.

Summarization, translation, Q&A, text generation and more at blazing speed using a T5 version implemented in ONNX. This package is still in alpha stag

Abel 211 Dec 28, 2022
RoNER is a Named Entity Recognition model based on a pre-trained BERT transformer model trained on RONECv2

RoNER RoNER is a Named Entity Recognition model based on a pre-trained BERT transformer model trained on RONECv2. It is meant to be an easy to use, hi

Stefan Dumitrescu 9 Nov 07, 2022
voice2json is a collection of command-line tools for offline speech/intent recognition on Linux

Command-line tools for speech and intent recognition on Linux

Michael Hansen 988 Jan 04, 2023
Code for Emergent Translation in Multi-Agent Communication

Emergent Translation in Multi-Agent Communication PyTorch implementation of the models described in the paper Emergent Translation in Multi-Agent Comm

Facebook Research 75 Jul 15, 2022
L3Cube-MahaCorpus a Marathi monolingual data set scraped from different internet sources.

L3Cube-MahaCorpus L3Cube-MahaCorpus a Marathi monolingual data set scraped from different internet sources. We expand the existing Marathi monolingual

21 Dec 17, 2022
Repository of the Code to Chatbots, developed in Python

Description In this repository you will find the Code to my Chatbots, developed in Python. I'll explain the structure of this Repository later. Requir

Li-am K. 0 Oct 25, 2022
In this project, we aim to achieve the task of predicting emojis from tweets. We aim to investigate the relationship between words and emojis.

Making Emojis More Predictable by Karan Abrol, Karanjot Singh and Pritish Wadhwa, Natural Language Processing (CSE546) under the guidance of Dr. Shad

Karanjot Singh 2 Jan 17, 2022
🌐 Translation microservice powered by AI

Dot Translate 🌐 A microservice for quick and local translation using A.I. This service starts a local webserver used for neural machine translation.

Dot HQ 48 Nov 22, 2022
Python SDK for working with Voicegain Speech-to-Text

Voicegain Speech-to-Text Python SDK Python SDK for the Voicegain Speech-to-Text API. This API allows for large vocabulary speech-to-text transcription

Voicegain 3 Dec 14, 2022
CJK computer science terms comparison / 中日韓電腦科學術語對照 / 日中韓のコンピュータ科学の用語対照 / 한·중·일 전산학 용어 대조

CJK computer science terms comparison This repository contains the source code of the website. You can see the website from the following link: Englis

Hong Minhee (洪 民憙) 88 Dec 23, 2022
Journey is a NLP-Powered Developer assistant

Journey Journey is a NLP-Powered Developer assistant Using on the powerful Natural Language Processing library Mindmeld, this projects aims to assist

Christian Eilers 21 Dec 11, 2022
Simple Annotated implementation of GPT-NeoX in PyTorch

Simple Annotated implementation of GPT-NeoX in PyTorch This is a simpler implementation of GPT-NeoX in PyTorch. We have taken out several optimization

labml.ai 101 Dec 03, 2022
KoBERTopic은 BERTopic을 한국어 데이터에 적용할 수 있도록 토크나이저와 BERT를 수정한 코드입니다.

KoBERTopic 모델 소개 KoBERTopic은 BERTopic을 한국어 데이터에 적용할 수 있도록 토크나이저와 BERT를 수정했습니다. 기존 BERTopic : https://github.com/MaartenGr/BERTopic/tree/05a6790b21009d

Won Joon Yoo 26 Jan 03, 2023
✨Fast Coreference Resolution in spaCy with Neural Networks

✨ NeuralCoref 4.0: Coreference Resolution in spaCy with Neural Networks. NeuralCoref is a pipeline extension for spaCy 2.1+ which annotates and resolv

Hugging Face 2.6k Jan 04, 2023
This is an incredibly powerful calculator that is capable of many useful day-to-day functions.

Description 💻 This is an incredibly powerful calculator that is capable of many useful day-to-day functions. Such functions include solving basic ari

Jordan Leich 37 Nov 19, 2022
Python module (C extension and plain python) implementing Aho-Corasick algorithm

pyahocorasick pyahocorasick is a fast and memory efficient library for exact or approximate multi-pattern string search meaning that you can find mult

Wojciech Muła 763 Dec 27, 2022
Maha is a text processing library specially developed to deal with Arabic text.

An Arabic text processing library intended for use in NLP applications Maha is a text processing library specially developed to deal with Arabic text.

Mohammad Al-Fetyani 184 Nov 27, 2022
Word2Wave: a framework for generating short audio samples from a text prompt using WaveGAN and COALA.

Word2Wave is a simple method for text-controlled GAN audio generation. You can either follow the setup instructions below and use the source code and CLI provided in this repo or you can have a play

Ilaria Manco 91 Dec 23, 2022