Deduplicating Training Data Makes Language Models Better

Overview

Deduplicating Training Data Makes Language Models Better

This repository contains code to deduplicate language model datasets as descrbed in the paper "Deduplicating Training Data Makes Language Models Better" by Katherine Lee, Daphne Ippolito, Andrew Nystrom, Chiyuan Zhang, Douglas Eck, Chris Callison-Burch and Nicholas Carlini. This repository contains both the ExactSubstr deduplication implementation (written in Rust) along with the scripts we used in the paper to perform deduplication and inspect the results (written in Python). In an upcoming update, we will add files to reproduce the NearDup-deduplicated versions of the C4, RealNews, LM1B, and Wiki-40B-en datasets.

This is not an officially supported Google product.

Why deduplicate?

When datasets are created by scraping raw text from the Internet, this will often result in the same sequences being repeated multiple times (e.g., we find a single 50 word sequence that is repeated in the C4 dataset 60,000 times). Training models on deduplicated datasets is faster (because they see fewer total examples) and experimentally results in models with similar or better perplexity to models trained on data that hasn't been deduplicated. Moreover, language models are less likely to exhibit memorization when their training data has been well-deduplicated.

Citing this work

If you use this repository or our deduplicated datasets you can cite

@article{lee2021deduplicating,
      title={Deduplicating Training Data Makes Language Models Better}, 
      author={Katherine Lee and Daphne Ippolito and Andrew Nystrom and Chiyuan Zhang and Douglas Eck and Chris Callison-Burch and Nicholas Carlini},
      journal={arXiv preprint arXiv:2107.06499},
      year={2021},
}

Exact Deduplication Code

We provide an implementation of the exact deduplication technique used in the paper. This is very much research code. It is (a very slightly cleaned up) version of exactly what we do in the paper. It assumes that you want to deduplicate something the size of C4 (~300GB) running on a machine with 96 cores and >600GB of RAM. If you only want to use this for reasonably-sized datasets, you should change the number of parallel threads from 96 to something smaller. If your machine is big enough, there should be no upper bound on the size of the dataset it can handle (well, 2^64-1 bytes is the limit, but I think we can all agree that's essentially unlimited).

We build a suffix array (based on Andrew Gallant's suffix array implementation) in src/table.rs. It has some minor changes from the original version that make it so we can't just import this library as a crate. First, we need 64-bit integers. The original implementation says that u32 works for "reasonably sized documents (~4GB)" but we're working with unreasonably sized documents. So we need u64. Second, we don't want UTF8 strings. Everything is a [u8] byte array, because we might be working over token sequences which aren't valid UTF8. The main complication in the rest of src/main.rs is the fact that we want things to run in parallel, and we probably can't fit the entire suffix array into memory. And so all of our algorithms are designed around these constraints.

If you just want to run the rust deduplicator, then you will only need to install Rust:

curl --proto '=https' --tlsv1.2 -sSf https://sh.rustup.rs | sh

If you additionally want to generate datasets to run the rust script on (and you probably do) then you will need python dependencies:

pip3 install numpy scipy tensorflow tensorflow_datasets transformers sentencepiece

Basic Usage

If you just want to reproduce the result of this paper, or deduplicate any language model that's already in the Tensorflow Datasets (TFDS) format, then you can just run the following commands:

cargo build

to compile the rust code, and then run

python3 scripts/load_dataset.py --data_dir $LOAD_DIR --save_dir $SAVE_DIR --name $DATASET --split $SPLIT [--tokenize]

For example, to get the LM1B training set you could run python3 scripts/load_dataset.py --data_dir ~/tensorflow_datasets --save_dir data --name lm1b --split test. This should will take just a few seconds to run on the test set or about an hour if running with the train set instead.

If the dataset is really big, you might want to add the --tokenize flag. This will shrink the dataset by roughly a factor of two by tokenizing it with the GPT-2 tokenizer.

And then to construct the suffix array run

python3 scripts/make_suffix_array.py [path/to/dataset]

For example, if you run python3 scripts/make_suffix_array.py data/lm1b.test, this will create a file data/lm1b.test.table.bin containing the suffix array. Again, this should be fast, about two hours on the LM1B train set when run single-thread and a few minutes on 96 cores.

(If you get an error that you have too many open files, that's because this script opens lots of files. You should run ulimit -Sn 1000000 to "fix" the error.)

Querying a suffix array to find duplicated examples

Start by loading and building a suffix array for a dataset as described above

Once you have the suffix array, you now query the dataset to find all occurances of a particular string. To do this, run

python3 scripts/count_occurances.py --suffix [path/to/suffix_array] [--query query_string] [--query_file /path/to/query]

On the LM1B test set, running python3 scripts/count_occurances.py --suffix data/lm1b.test --query " on Tuesday" should return 1288. If you tokenized the dataset, then you should pass --tokenizetocount_occurences.py` as well, to get the same result (plus or minus tokenization differences).

If you want to confirm this the outputted number is correct (assuming you haven't tokenized), you can run cat /tmp/lm1b.test | grep -ao " on Tuesday" and get the same result.

Advanced Usage

The above scripts work by calling into the core Rust suffix array deduplicator. If you want to do each step yourself, the following options are available:

Single threaded suffix array construction

To build a suffix array for any particular file, you can run

cargo run save [file_path]

This will create a file called [file_path].table.bin which contains the suffix array for the file provided. This algorithm is linear time, but (a) only runs on a single core, and (b) has memory requirement O(big * len(file)) which is prohibitive for large files.

Parallel suffix array construction

To build a suffix array for an extremely large file (e.g., ~about as much RAM as available) it is better to run the script

python scripts/make_suffix_array.py [file_path]

This script will build the suffix array in parallel by splitting the single file into chunks, generating suffix arrays for each chunk, and then merging the suffix arrays together to form the full suffix array. Note that in general this algorithm is quadratic, but when the maximum substring length is short relative to the total file length (as it is, when generating suffix arrays for N independent training examples) it will never reach this worst case behavior.

The two steps are described below.

Building a piece of a suffix array from a piece of a file

The first generats a suffix array from a piece of a file. This is implemented by running

cargo run save_part [file_path] [byte_start] [byte_end]

And builds a suffix array for the byte sequence between [byte_start] and [byte_end] for the given file. Multiple of these can be run in parallel to build a suffix array for a file quickly.

Merging suffix array pieces to create a single suffix array

Given the several independent suffix arrays, merging them is now just a matter of calling

cargo run merge_parallel [path_to_partial_suffix_trees,...] [tmp_output_directory]

to generate a collection of ordered suffix arrays pieces in the output directory. The final step just requires merging these together

cat [tmp_output_directory]/* > [file_path].table.bin

Finding Duplicates

Given a suffix array file, as generated in the prevous section, it can now be queried for interesting statistics. The simplest operation, counting occurrences of particular substrings, takes O(log(N)) time and O(query_length) memory requirements, (as shown above with scripts/count_occurances.py). To do this you can run:

cargo run count_occurances /path/to/dataset /path/to/query_file

(Indeed, the python script is just a wrapper that makes calling this nicer, with the option for tokenization.) This is useful mainly as a commandline interface to interact with the dataset to find interesting properties. To run more sophisticated analysis, use the tools described below:

Finding duplicates between two documents

Given a document A and another document B, we can find all duplicates betwen the two by (1) constructing suffix arrays for both, and then (2) linearly walking the suffix arrays in order to find all duplicates of a given length.

Once the suffix array for the dataset has been constructed, this algorithm therefore requires time O(len(dataset) + len(query)) and space O(len(dataset)). It is better to run this algorithm when the number of queries into the dataset is greater than O(len(dataset)/log(len(query))). However note that the prior code requires disk seeks and and this implementation is a linear scan through the suffix array table, so in practice there is at least a factor-of-10 speedup here. As a rough order of magnitude, for a dataset with ~100GB, it is faster to run similar_parallel when querying with more than a few megabytes of text. Otherwise it is probably faster to run count_occurances.

Notice that this command also requires that the entire dataset fits in memory. For many datasets this is not a problem, but the C4 dataset is 350 GB and the Pile dataset is 750 GB (both even after tokenization). The machine must therefore have a lot of RAM for this to work.

cargo run similar_parallel [dataset1] [dataset2]

This creates lots of containing the position of all examples in dataset2 that are also in dataset1. (The code could also do the inverse at the same time, if you want to modify it slightly.) However it spits this out in some not-very-useful form: a list of tokens x_i so that dataset2[x_i:x_i+100] is also in dataset1. But this probably has overlaps.

The second step is then to run

cargo run collect_similar [dataset2]. This converts the result to instead compute ranges so that instead we have dataset2[xi:yi] match.

Finding duplicates within one document

To find duplicates that are contained within one document (for example, to actually deduplicate a dataset as we do in the paper) run the command

cargo run selfsimilar_parallel [dataset]

This will find all repeated substrings contained in the dataset above a given length threshold. Again run collect_similar to find the indexs of repeated examples.

Approx Deduplication Results

Coming soon.

Owner
Google Research
Google Research
This package contains deep learning models and related scripts for RoseTTAFold

RoseTTAFold This package contains deep learning models and related scripts to run RoseTTAFold This repository is the official implementation of RoseTT

1.6k Jan 03, 2023
A tutorial on DataFrames.jl prepared for JuliaCon2021

JuliaCon2021 DataFrames.jl Tutorial This is a tutorial on DataFrames.jl prepared for JuliaCon2021. A video recording of the tutorial is available here

Bogumił Kamiński 106 Jan 09, 2023
Augmenting Physical Models with Deep Networks for Complex Dynamics Forecasting

Official code of APHYNITY Augmenting Physical Models with Deep Networks for Complex Dynamics Forecasting (ICLR 2021, Oral) Yuan Yin*, Vincent Le Guen*

Yuan Yin 24 Oct 24, 2022
PyTorch implementation for MINE: Continuous-Depth MPI with Neural Radiance Fields

MINE: Continuous-Depth MPI with Neural Radiance Fields Project Page | Video PyTorch implementation for our ICCV 2021 paper. MINE: Towards Continuous D

Zijian Feng 325 Dec 29, 2022
Unimodal Face Classification with Multimodal Training

Unimodal Face Classification with Multimodal Training This is a PyTorch implementation of the following paper: Unimodal Face Classification with Multi

Wenbin Teng 3 Jul 06, 2022
Style transfer between images was performed using the VGG19 model

Style transfer between images was performed using the VGG19 model. The necessary codes, libraries and all other information of this project are available below

Onur yılmaz 2 May 09, 2022
[ICLR 2022 Oral] F8Net: Fixed-Point 8-bit Only Multiplication for Network Quantization

F8Net Fixed-Point 8-bit Only Multiplication for Network Quantization (ICLR 2022 Oral) OpenReview | arXiv | PDF | Model Zoo | BibTex PyTorch implementa

Snap Research 76 Dec 13, 2022
PyTorch implementation for paper StARformer: Transformer with State-Action-Reward Representations.

StARformer This repository contains the PyTorch implementation for our paper titled StARformer: Transformer with State-Action-Reward Representations.

Jinghuan Shang 14 Dec 09, 2022
Repo público onde postarei meus estudos de Python, buscando aprender por meio do compartilhamento do aprendizado!

Seja bem vindo à minha repo de Estudos em Python 3! Este é um repositório criado por um programador amador que estuda tópicos de finanças, estatística

32 Dec 24, 2022
[NeurIPS 2020] Semi-Supervision (Unlabeled Data) & Self-Supervision Improve Class-Imbalanced / Long-Tailed Learning

Rethinking the Value of Labels for Improving Class-Imbalanced Learning This repository contains the implementation code for paper: Rethinking the Valu

Yuzhe Yang 656 Dec 28, 2022
HybridNets: End-to-End Perception Network

HybridNets: End2End Perception Network HybridNets Network Architecture. HybridNets: End-to-End Perception Network by Dat Vu, Bao Ngo, Hung Phan 📧 FPT

Thanh Dat Vu 370 Dec 29, 2022
Implementation of Advantage-Weighted Regression: Simple and Scalable Off-Policy Reinforcement Learning

advantage-weighted-regression Implementation of Advantage-Weighted Regression: Simple and Scalable Off-Policy Reinforcement Learning, by Peng et al. (

Omar D. Domingues 1 Dec 02, 2021
Implementation of "Generalizable Neural Performer: Learning Robust Radiance Fields for Human Novel View Synthesis"

Generalizable Neural Performer: Learning Robust Radiance Fields for Human Novel View Synthesis Abstract: This work targets at using a general deep lea

163 Dec 14, 2022
PyTorch implementation of Histogram Layers from DeepHist: Differentiable Joint and Color Histogram Layers for Image-to-Image Translation

deep-hist PyTorch implementation of Histogram Layers from DeepHist: Differentiable Joint and Color Histogram Layers for Image-to-Image Translation PyT

Winfried Lötzsch 10 Dec 06, 2022
Source code for deep symbolic optimization.

Update July 10, 2021: This repository now supports an additional symbolic optimization task: learning symbolic policies for reinforcement learning. Th

Brenden Petersen 290 Dec 25, 2022
Align before Fuse: Vision and Language Representation Learning with Momentum Distillation

This is the official PyTorch implementation of the ALBEF paper [Blog]. This repository supports pre-training on custom datasets, as well as finetuning on VQA, SNLI-VE, NLVR2, Image-Text Retrieval on

Salesforce 805 Jan 09, 2023
PyTorch-Multi-Style-Transfer - Neural Style and MSG-Net

PyTorch-Style-Transfer This repo provides PyTorch Implementation of MSG-Net (ours) and Neural Style (Gatys et al. CVPR 2016), which has been included

Hang Zhang 906 Jan 04, 2023
torchsummaryDynamic: support real FLOPs calculation of dynamic network or user-custom PyTorch ops

torchsummaryDynamic Improved tool of torchsummaryX. torchsummaryDynamic support real FLOPs calculation of dynamic network or user-custom PyTorch ops.

Bohong Chen 1 Jan 07, 2022
Syed Waqas Zamir 906 Dec 30, 2022
Data visualization app for H&M competition in kaggle

handm_data_visualize_app Data visualization app by streamlit for H&M competition in kaggle. competition page: https://www.kaggle.com/competitions/h-an

Kyohei Uto 12 Apr 30, 2022