Introduction

This repository contains the code for Instance-wise Ordered Transformer (IOT), which is introduced in the ICLR2021 paper IOT: Instance-wise Layer Reordering for Transformer Structures.

If you find this work helpful in your research, please cite as:

@inproceedings{
zhu2021iot,
title={{\{}IOT{\}}: Instance-wise Layer Reordering for Transformer Structures},
author={Jinhua Zhu and Lijun Wu and Yingce Xia and Shufang Xie and Tao Qin and Wengang Zhou and Houqiang Li and Tie-Yan Liu},
booktitle={International Conference on Learning Representations},
year={2021},
url={https://openreview.net/forum?id=ipUPfYxWZvM}
}

Requirements and Installation

PyTorch version == 1.0.0
Python version >= 3.5

To install IOT:

git clone https://github.com/instance-wise-ordered-transformer/IOT
cd IOT
pip install --editable .

Getting Started

Take IWSLT14 De-En translation as an example.

Data Preprocessing

cd examples/translation/
bash prepare-iwslt14.sh
cd ../..

TEXT=examples/translation/iwslt14.tokenized.de-en
python preprocess.py --source-lang de --target-lang en \
    --trainpref $TEXT/train --validpref $TEXT/valid --testpref $TEXT/test \
    --destdir data-bin/iwslt14.tokenized.de-en --joined-dictionary

Training

Encoder order is set to be the default one without reordering (ENCODER_MAX_ORDER=1), since the paper finds that both reordering encoder and decoder is not good as reordering decoder only.

#!/bin/bash
export CUDA_VISIBLE_DEVICES=${1:-0}
nvidia-smi

ENCODER_MAX_ORDER=1
DECODER_MAX_ORDER=3
DECODER_ORDER="0 3 5"
DIVERSITY=0.1
GS_MAX=20
GS_MIN=2
GS_R=0
GS_UF=5000
KL=0.01
CLAMPVAL=0.05

DECODER_ORDER_NAME=`echo $DECODER_ORDER | sed 's/ //g'`
SAVE_DIR=checkpoints/dec_${DECODER_MAX_ORDER}_order_${DECODER_ORDER_NAME}_div_${DIVERSITY}_gsmax_${GS_MAX}_gsmin_${GS_MIN}_gsr_${GS_R}_gsuf_${GS_UF}_kl_${KL}_clampval_${CLAMPVAL}
mkdir -p ${SAVE_DIR}

python -u train.py data-bin/iwslt14.tokenized.de-en -a transformer_iwslt_de_en \
--optimizer adam --lr 0.0005 -s de -t en --label-smoothing 0.1 --dropout 0.3 --max-tokens 4000 \
--min-lr 1e-09 --lr-scheduler inverse_sqrt --weight-decay 0.0001 --criterion label_smoothed_cross_entropy \
--max-update 100000 --warmup-updates 4000 --warmup-init-lr 1e-07 --adam-betas '(0.9,0.98)' \
--save-dir $SAVE_DIR --share-all-embeddings  --gs-clamp --decoder-orders $DECODER_ORDER  \
--encoder-max-order $ENCODER_MAX_ORDER  --decoder-max-order $DECODER_MAX_ORDER  --diversity $DIVERSITY \
--gumbel-softmax-max $GS_MAX  --gumbel-softmax-min $GS_MIN --gumbel-softmax-tau-r $GS_R  --gumbel-softmax-update-freq $GS_UF \
--kl $KL --clamp-value $CLAMPVAL | tee -a ${SAVE_DIR}/train.log

Evaluation

#!/bin/bash
set -x
set -e

pip install -e . --user
export CUDA_VISIBLE_DEVICES=${1:-0}
nvidia-smi

ENCODER_MAX_ORDER=1
DECODER_MAX_ORDER=3
DECODER_ORDER="0 3 5"
DIVERSITY=0.1
GS_MAX=20
GS_MIN=2
GS_R=0
GS_UF=5000
KL=0.01
CLAMPVAL=0.05

DECODER_ORDER_NAME=`echo $DECODER_ORDER | sed 's/ //g'`
SAVE_DIR=checkpoints/dec_${DECODER_MAX_ORDER}_order_${DECODER_ORDER_NAME}_div_${DIVERSITY}_gsmax_${GS_MAX}_gsmin_${GS_MIN}_gsr_${GS_R}_gsuf_${GS_UF}_kl_${KL}_clampval_${CLAMPVAL}

python generate.py data-bin/iwslt14.tokenized.de-en \
  --path $SAVE_DIR/checkpint_best.pt \
  --batch-size 128 --beam 5 --remove-bpe --quiet --num-ckts $DECODER_MAX_ORDER

IOT: Instance-wise Layer Reordering for Transformer Structures

Related tags

Overview

Introduction

Requirements and Installation

Getting Started

Data Preprocessing

Training

Evaluation

Owner

IOT

I tried to apply the CAM algorithm to YOLOv4 and it worked.

Reusable constraint types to use with typing.Annotated

This is a virtual picture dragging application. Users may virtually slide photos across the screen. The distance between the index and middle fingers determines the movement. Smaller distances indicate click and motion, whereas bigger distances indicate only hand movement.

Distributional Sliced-Wasserstein distance code

This code finds bounding box of a single human mouth.

Temporal Knowledge Graph Reasoning Triggered by Memories

Event sourced bank - A wide-and-shallow example using the Python event sourcing library

Code for "OctField: Hierarchical Implicit Functions for 3D Modeling (NeurIPS 2021)"

Attendance Monitoring with Face Recognition using Python

True Few-Shot Learning with Language Models

PyTorch implementation of the ExORL: Exploratory Data for Offline Reinforcement Learning

NeWT: Natural World Tasks

Code for ECCV 2020 paper "Contacts and Human Dynamics from Monocular Video".

Pyramid Pooling Transformer for Scene Understanding

Unofficial Alias-Free GAN implementation. Based on rosinality's version with expanded training and inference options.

Toolbox of models, callbacks, and datasets for AI/ML researchers.

Locally Enhanced Self-Attention: Rethinking Self-Attention as Local and Context Terms

PyTorch Implementation of Unsupervised Depth Completion with Calibrated Backprojection Layers (ORAL, ICCV 2021)

This is RFA-Toolbox, a simple and easy-to-use library that allows you to optimize your neural network architectures using receptive field analysis (RFA) and create graph visualizations of your architecture.

Manage the availability of workspaces within Frappe/ ERPNext (sidebar) based on user-roles