Model parallel transformers in JAX and Haiku

Overview

Table of contents

  1. Mesh Transformer JAX
    1. Updates
  2. Pretrained Models
    1. GPT-J-6B
      1. Links
      2. Acknowledgments
      3. License
      4. Model Details
      5. Zero-Shot Evaluations
  3. Architecture and Usage
    1. Fine-tuning
    2. JAX Dependency
  4. TODO

Mesh Transformer JAX

A haiku library using the xmap/pjit operators in JAX for model parallelism of transformers.

The parallelism scheme is similar to the original Megatron-LM, which is efficient on TPUs due to the high speed 2d mesh network. There is also an experimental model version which implements ZeRo style sharding.

This library is designed for scalability up to approximately 40B parameters on TPUv3s, beyond which different parallelism strategies should be used. See other implementations such as GPT-NeoX or DeepSpeed for that.

One future direction for research is integrating this codebase with swarm-jax, to achieve further scalability with pipeline parallelism.

Updates

12-07-21: Added guide to fine tuning

Pretrained Models

GPT-J-6B

A 6 billion parameter, autoregressive text generation model trained on The Pile.

Links

Slim weights (bf16 weights only, for inference, 9GB)

Full weights (including optimizer params, 61GB)

Colab demo

Web demo

Aran's blog post

Acknowledgments

This project would not have been possible without compute generously provided by the TPU Research Cloud with assistance from EleutherAI.

Thanks to the Cloud TPU team at Google for providing early access to the Cloud TPU VM alpha (now publicly available!)

Thanks to everyone who have helped out one way or another (listed alphabetically):

  • Aran Komatsuzaki for advice with experiment design and writing the blog posts.
  • James Bradbury for valuable assistance with debugging JAX issues.
  • Janko Prester for creating the web demo frontend.
  • Laurence Golding for adding some features to the web demo.
  • Leo Gao for running zero shot evaluations for the baseline models for the table.

License

The weights of GPT-J-6B are licensed under version 2.0 of the Apache License.

Model Details

Hyperparameter Value
n_parameters 6,053,381,344
n_layers 28*
d_model 4,096
d_ff 16,384
n_heads 16
d_head 256
n_ctx 2,048
n_vocab 50,257 (same tokenizer as GPT-2/3)
position encoding Rotary position encodings (RoPE)
RoPE dimensions 64

* each layer consists of one feedforward block and one self attention block

The model consists of 28 layers with a model dimension of 4096, and a feedforward dimension of 16384. The model dimension is split into 16 heads, each with a dimension of 256. Rotary position encodings (RoPE) was applied to 64 dimensions of each head. The model is trained with a tokenization vocabulary of 50257, using the same set of BPEs as GPT-2/GPT-3.

Zero-Shot Evaluations

Models roughly sorted by performance, or by FLOPs if not available.

Model Weights Training FLOPs LAMBADA PPL ↓ LAMBADA Acc ↑ Winogrande ↑ Hellaswag ↑ PIQA ↑ Dataset Size (GB)
Chance 0 ~a lot ~0% 50% 25% 25% 0
GPT-3-Ada‡ ----- 9.95 51.6% 52.9% 43.4% 70.5% -----
GPT-2-1.5B ----- 10.63 51.21% 59.4% 50.9% 70.8% 40
GPTNeo-1.3B‡ 3.0e21 7.50 57.2% 55.0% 48.9% 71.1% 825
Megatron-2.5B* 2.4e21 ----- 61.7% ----- ----- ----- 174
GPTNeo-2.7B‡ 6.8e21 5.63 62.2% 56.5% 55.8% 73.0% 825
GPT-3-1.3B*‡ 2.4e21 5.44 63.6% 58.7% 54.7% 75.1% ~800
GPT-3-Babbage‡ ----- 5.58 62.4% 59.0% 54.5% 75.5% -----
Megatron-8.3B* 7.8e21 ----- 66.5% ----- ----- ----- 174
GPT-3-2.7B*‡ 4.8e21 4.60 67.1% 62.3% 62.8% 75.6% ~800
Megatron-11B† 1.0e22 ----- ----- ----- ----- ----- 161
GPT-J-6B 1.5e22 3.99 69.7% 65.3% 66.1% 76.5% 825
GPT-3-6.7B*‡ 1.2e22 4.00 70.3% 64.5% 67.4% 78.0% ~800
GPT-3-Curie‡ ----- 4.00 69.3% 65.6% 68.5% 77.9% -----
GPT-3-13B*‡ 2.3e22 3.56 72.5% 67.9% 70.9% 78.5% ~800
GPT-3-175B*‡ 3.1e23 3.00 76.2% 70.2% 78.9% 81.0% ~800
GPT-3-Davinci‡ ----- 3.0 75% 72% 78% 80% -----
Gopher 230B* 6.31E+23 ----- 74.50% 70.10% 79.20% 81.80% 1344
MT-NLG 530B*‡ ----- ----- 76.6% 73.0% 80.2% 82.0% -----

* represents evaluation numbers reported by their respective authors, all other numbers are provided by running the lm-evaluation-harness either with the released weights or with API access. Due to subtle implementation differences as well as different zero shot task framing, these might not be directly comparable. See this blog post for more details.

The Megatron-11B model provides no comparable metrics, and several implementations using the released weights do not reproduce the generation quality and evaluations. (see 1 2 3) Thus, evaluation was not attempted.

These models have been trained with data which contains possible test set contamination. The OpenAI GPT-3 models failed to deduplicate training data for certain test sets, while the GPT-Neo models as well as this one is trained on The Pile, which has not been deduplicated against any test sets.

Architecture and Usage

Most scripts in this repository are designed to be run on TPUs, which under the TPU-VM architecture are virtual machines which can run arbitrary code. Most scripts are designed to spin up a TPU, SSH into it to set up the dependencies and copy code over from the local directory, and then start a Ray worker which can accept RPC calls.

The TPUVMs handles running model training steps and evaluation, checkpoint save and loading, while the driver python program handles data loading and general orchestration (such as when to save checkpoints etc).

This means that most scripts (train.py, eval_harness.py etc) expect to be running on a GCE virtual machine in the same region as the TPUs, to minimize RPC latency and data transfer cost. Other scripts (usually ones which don't take a --tpu argument, such as device_sample.py, device_serve.py or device_train.py) expect to be run directly on a TPUVM. The device_* scripts only work on a v3-8 and not on larger pods.

Furthermore, there is an example (resharding_example.py) of how to convert the provided checkpoints (which have 8 shards in the case of GPT-J-6B) down to a smaller number, such as for when running on GPU(s).

Fine-tuning

To fine-tune the model, run device_train.py on a TPU VM. Using a TPU v3-8, you can fine-tune at a rate of ~5000 tokens/second, which should be sufficient for small-to-medium-size datasets.

Please read the step by step guide for thorough fine-tuning instructions.

JAX Dependency

Note this library has some specific requirements for JAX version. Specifically, to use the v1 models (including GPT-J 6B), jax==0.2.12 is required. This in turn depends on jaxlib==0.1.68. If this is not done, you will get cryptic xmap errors

However, to use the v2 model code (no publicly released weights), the newest JAX version can be used.

Citation

To cite this repository:

@misc{mesh-transformer-jax,
  author = {Wang, Ben},
  title = {{Mesh-Transformer-JAX: Model-Parallel Implementation of Transformer Language Model with JAX}},
  howpublished = {\url{https://github.com/kingoflolz/mesh-transformer-jax}},
  year = 2021,
  month = May
}

To cite the weights of GPT-J-6B:

@misc{gpt-j,
  author = {Wang, Ben and Komatsuzaki, Aran},
  title = {{GPT-J-6B: A 6 Billion Parameter Autoregressive Language Model}},
  howpublished = {\url{https://github.com/kingoflolz/mesh-transformer-jax}},
  year = 2021,
  month = May
}

If you use this repository or any of the pretrained weights to do something cool, we would love to hear about it. Feel free to open a github issue or reach out over email (in profile).

TODO

  • disentangle heads and shards
  • test/benchmark on TPU
  • implement gradient checkpointing
  • fix initialization
  • mixed precision
  • deal with preemptible TPUs
  • test and validate generation
  • shard activations instead of replicating for memory efficiency (in v2)
  • support ZeRO style sharding (in v2)
Owner
Ben Wang
Ben Wang
Pretrained Japanese BERT models

Pretrained Japanese BERT models This is a repository of pretrained Japanese BERT models. The models are available in Transformers by Hugging Face. Mod

Inui Laboratory 387 Dec 30, 2022
Repository for the paper: VoiceMe: Personalized voice generation in TTS

🗣 VoiceMe: Personalized voice generation in TTS Abstract Novel text-to-speech systems can generate entirely new voices that were not seen during trai

Pol van Rijn 80 Dec 29, 2022
Python code for ICLR 2022 spotlight paper EViT: Expediting Vision Transformers via Token Reorganizations

Expediting Vision Transformers via Token Reorganizations This repository contain

Youwei Liang 101 Dec 26, 2022
ChainKnowledgeGraph, 产业链知识图谱包括A股上市公司、行业和产品共3类实体

ChainKnowledgeGraph, 产业链知识图谱包括A股上市公司、行业和产品共3类实体,包括上市公司所属行业关系、行业上级关系、产品上游原材料关系、产品下游产品关系、公司主营产品、产品小类共6大类。 上市公司4,654家,行业511个,产品95,559条、上游材料56,824条,上级行业480条,下游产品390条,产品小类52,937条,所属行业3,946条。

liuhuanyong 415 Jan 06, 2023
Code for our paper "Mask-Align: Self-Supervised Neural Word Alignment" in ACL 2021

Mask-Align: Self-Supervised Neural Word Alignment This is the implementation of our work Mask-Align: Self-Supervised Neural Word Alignment. @inproceed

THUNLP-MT 46 Dec 15, 2022
NLP - Machine learning

Flipkart-product-reviews NLP - Machine learning About Product reviews is an essential part of an online store like Flipkart’s branding and marketing.

Harshith VH 1 Oct 29, 2021
An ActivityWatch watcher to pose questions to the user and record her answers.

aw-watcher-ask An ActivityWatch watcher to pose questions to the user and record her answers. This watcher uses Zenity to present dialog boxes to the

Bernardo Chrispim Baron 33 Dec 03, 2022
FewCLUE: 为中文NLP定制的小样本学习测评基准

FewCLUE: 为中文NLP定制的小样本学习测评基准

CLUE benchmark 387 Jan 04, 2023
Automated question generation and question answering from Turkish texts using text-to-text transformers

Turkish Question Generation Offical source code for "Automated question generation & question answering from Turkish texts using text-to-text transfor

Open Business Software Solutions 29 Dec 14, 2022
The repository for the paper: Multilingual Translation via Grafting Pre-trained Language Models

Graformer The repository for the paper: Multilingual Translation via Grafting Pre-trained Language Models Graformer (also named BridgeTransformer in t

22 Dec 14, 2022
Easy to start. Use deep nerual network to predict the sentiment of movie review.

Easy to start. Use deep nerual network to predict the sentiment of movie review. Various methods, word2vec, tf-idf and df to generate text vectors. Various models including lstm and cov1d. Achieve f1

1 Nov 19, 2021
BARTpho: Pre-trained Sequence-to-Sequence Models for Vietnamese

Table of contents Introduction Using BARTpho with fairseq Using BARTpho with transformers Notes BARTpho: Pre-trained Sequence-to-Sequence Models for V

VinAI Research 58 Dec 23, 2022
Adversarial Examples for Extreme Multilabel Text Classification

Adversarial Examples for Extreme Multilabel Text Classification The code is adapted from the source codes of BERT-ATTACK [1], APLC_XLNet [2], and Atte

1 May 14, 2022
Material for GW4SHM workshop, 16/03/2022.

GW4SHM Workshop Wednesday, 16th March 2022 (13:00 – 15:15 GMT): Presented by: Dr. Rhodri Nelson, Imperial College London Project website: https://www.

Devito Codes 1 Mar 16, 2022
Open-Source Toolkit for End-to-End Speech Recognition leveraging PyTorch-Lightning and Hydra.

OpenSpeech provides reference implementations of various ASR modeling papers and three languages recipe to perform tasks on automatic speech recogniti

Soohwan Kim 26 Dec 14, 2022
[ICLR'19] Trellis Networks for Sequence Modeling

TrellisNet for Sequence Modeling This repository contains the experiments done in paper Trellis Networks for Sequence Modeling by Shaojie Bai, J. Zico

CMU Locus Lab 460 Oct 13, 2022
Simple Python script to scrape youtube channles of "Parity Technologies and Web3 Foundation" and translate them to well-known braille language or any language

Simple Python script to scrape youtube channles of "Parity Technologies and Web3 Foundation" and translate them to well-known braille language or any

Little Endian 1 Apr 28, 2022
TFIDF-based QA system for AIO2 competition

AIO2 TF-IDF Baseline This is a very simple question answering system, which is developed as a lightweight baseline for AIO2 competition. In the traini

Masatoshi Suzuki 4 Feb 19, 2022
This code is the implementation of Text Emotion Recognition (TER) with linguistic features

APSIPA-TER This code is the implementation of Text Emotion Recognition (TER) with linguistic features. The network model is BERT with a pretrained mod

kenro515 1 Feb 08, 2022
Production First and Production Ready End-to-End Keyword Spotting Toolkit

Production First and Production Ready End-to-End Keyword Spotting Toolkit

223 Jan 02, 2023