fighterhit / EET

Easy and Efficient Transformer

中文README

EET(Easy and Efficient Transformer) is an efficient Pytorch inference plugin focus on Transformer-based models with large model sizes and long sequences.

Features

1、Joint-Decoding.  Propose and implement a novel decoding mechanism, an optimization for incremental decoding, increasing the token-parallelism and improving GPU occupancy.
2、High performance. Design highly optimized CUDA kernels, referencing to NVIDIA Faster Transformer, supporting long sequences as well as large model sizes with advanced optimization.
3、Flexible.  Provide op-level and model-level APIs.
4、Easy to use. EET could be integrated into Fairseq and Transformes directly.
5、Smart deployment. Support dynamic batching and variable input length, Combined with python web, EET could be deployed smoothly.

EET has been applied to a variety of NetEase online services. In the future, EET will work on urtra-large-scale model inference of trillion parameters.

• A novel Joint-Decoding mechanism
• Quick Start
  • Environment
  • Installation
  • Run
    • run BERT in Transformers
    • run GPT2 in Transformers
    • run GPT2 in Fairseq
• Supported Models
  • BERT
  • GPT2
• Usage
• APIs
• Performance
  • We show GPT2 inference performance here.
  • We show BERT inference performance here.
• TODO
• Contact us

A novel Joint-Decoding mechanism

Three-level decoding:

• Level-1. Joint prompts information insides a sequence, handled in parallel.

• Level-2. Joint prompts information cross whole batch, along with left-padding to ensure the correctness of inferred results.

• Level-3. Joint Full-decoding with Incremental-decoding towards the best performance.

Quick Start

Environment

• cuda:>=10.1
• python:>=3.7
• gcc:>= 7.4.0
• torch:>=1.5.0
• numpy:>=1.19.1

Installation

From Source

If you are installing from source, you will need install the necessary environment.Then, proceed as follows:

$ git clone [e-mail]:NetEase-FuXi/EET.git
$ pip install transformers==3.0.2
$ pip installl fairseq==0.10.0
$ pip install .

Due to the compilation of a large number of cuda kernels, the installation time is relatively long, please be patient.

From Docker

$ git clone [e-mail]:NetEase-FuXi/EET.git
$ cd EET/docker
$ docker build -t your_docker_name:your_docker_version .
$ nvidia-docker run -it --net=host -v /your/project/directory/:/root/workspace  Your_Docker_Name:Your_docker_version bash

EET has been installed in the docker.

Run

run BERT in Transformers

$ cd EET/example  
$ python bert_transformers_example.py

run GPT2 in Transformers

$ cd EET/example    
$ python gpt2_transformers_example.py

run GPT2 in Fairseq

$ cd EET/example    
$ python gpt2_fairseq_example.py

Supported Models

We currenly support the GPT-2, Bert.

GPT2

BERT

Usage

EET provides python User-friendly APIs(python/eet), integrated into Fairseq and Transformers with just a few lines of code.

1、How to inference

2、How to customize model
You can refer to Operators APIs listed below to build your own model structure, just by modifying the files under python/eet.

3、How to integrate EET into fairseq
Replace the original transformer.py in Fairseq with our transformer.py and reinstall the Fairseq, that is all ! Transformer.py in EET corresponds to the fusion of transformer.py and transformer_layer.py in fairseq.

4、How to integrate EET into Transformers Replace the original modeling_bert.py and modeling_gpt2.py in Transformers with our modeling_bert.py and modeling_gpt2.py and reinstall the Transformers, that is all ! modeling_bert.py in EET corresponds to modeling_bert.py in transformers;modeling_gpt2.py in EET corresponds to modelling_gpt2.py in transformers.

5、How to make a server
We choose service-streamer to make the model server, building the service based on your python project directly. Please make sure the dynamic-batch is open if you want a higher throughput.

APIs

1. model APIs:We provide ready-made APIs for GPT2 and Bert models.

   EET and fairseq class comparison table

   EET	fairseq	Remarks
   EETTransformerDecoder	TransformerDecoder
   EETTransformerDecoderLayer	TransformerDecoderLayer
   EETTransformerAttention	MultiheadAttention
   EETTransformerFeedforward	TransformerDecoderLayer	fusion of multiple small operators
   EETTransformerEmbedding	Embedding + PositionalEmbedding
   EETTransformerLayerNorm	nn.LayerNorm

   EET and transformers class comparison table

   EET	transformers	Remarks
   EETBertModel	BertModel
   EETBertEncoder	BertEncoder
   EETBertEncoderLayer	BertLayer
   EETBertAttention	BertAttention
   EETBertFeedforward	BertIntermediate + BertOutput
   EETBertEmbedding	BertEmbeddings
   EETGPT2Model	GPT2Model
   EETGPT2Decoder	GPT2Model	transformers has no GPT2Decoder
   EETGPT2DecoderLayer	Block
   EETGPT2Attention	Attention
   EETGPT2Feedforward	MLP
   EETGPT2Embedding	nn.Embedding
   EETLayerNorm	nn.LayerNorm

2. operators APIs:We provide all the operators required for Transformer models. You can combine different kernels to build different model structures

   operators APIs	Remarks
   masked_multi_head_attention	GPT2 self_attention
   cross_multi_head_attention	cross_attention
   multi_head_attention	Bert self_attention
   ffn	FeedForwardNetwork
   embedding	transformers & fairseq
   layernorm	nn.LayerNorm

Performance

We tested the performance of EET on two GPU hardware platforms. We chose pytorch, NVIDIA Faster Transformers, and lightseq implementations for comparison. For the results of the experiment, please click on the link below.

benchmark

TODO

1. int8
2. sparse

Contact us

You can post your problem with github issues.