ActNN: Reducing Training Memory Footprint via 2-Bit Activation Compressed Training

Related tags

Deep Learningactnn
Overview

ActNN : Activation Compressed Training

This is the official project repository for ActNN: Reducing Training Memory Footprint via 2-Bit Activation Compressed Training by Jianfei Chen*, Lianmin Zheng*, Zhewei Yao, Dequan Wang, Ion Stoica, Michael W. Mahoney, and Joseph E. Gonzalez.

TL; DR. ActNN is a PyTorch library for memory-efficient training. It reduces the training memory footprint by compressing the saved activations. ActNN is implemented as a collection of memory-saving layers. These layers have an identical interface to their PyTorch counterparts.

Abstract

The increasing size of neural network models has been critical for improvements in their accuracy, but device memory is not growing at the same rate. This creates fundamental challenges for training neural networks within limited memory environments. In this work, we propose ActNN, a memory-efficient training framework that stores randomly quantized activations for back propagation. We prove the convergence of ActNN for general network architectures, and we characterize the impact of quantization on the convergence via an exact expression for the gradient variance. Using our theory, we propose novel mixed-precision quantization strategies that exploit the activation's heterogeneity across feature dimensions, samples, and layers. These techniques can be readily applied to existing dynamic graph frameworks, such as PyTorch, simply by substituting the layers. We evaluate ActNN on mainstream computer vision models for classification, detection, and segmentation tasks. On all these tasks, ActNN compresses the activation to 2 bits on average, with negligible accuracy loss. ActNN reduces the memory footprint of the activation by 12×, and it enables training with a 6.6× to 14× larger batch size.

mem_speed_r50 Batch size vs. training throughput on ResNet-50. Red cross mark means out-of-memory. The shaded yellow region denotes the possible batch sizes with full precision training. ActNN achieves significantly larger maximum batch size over other state-of-the-art systems and displays a nontrivial trade-off curve.

Install

  • Requirements
torch>=1.7.1
torchvision>=0.8.2
  • Build
cd actnn
pip install -v -e .

Usage

mem_speed_benchmark/train.py is an example on using ActNN for models from torchvision.

Basic Usage

  • Step1: Configure the optimization level
    ActNN provides several optimization levels to control the trade-off between memory saving and computational overhead. You can set the optimization level by
import actnn
# available choices are ["L0", "L1", "L2", "L3", "L4", "L5"]
actnn.set_optimization_level("L3")

See set_optimization_level for more details.

  • Step2: Convert the model to use ActNN's layers.
model = actnn.QModule(model)

Note:

  1. Convert the model before calling .cuda().
  2. Set the optimization level before invoking actnn.QModule or constructing any ActNN layers.
  3. Automatic model conversion only works with standard PyTorch layers. Please use the modules (nn.Conv2d, nn.ReLU, etc.), not the functions (F.conv2d, F.relu).
  • Step3: Print the model to confirm that all the modules (Conv2d, ReLU, BatchNorm) are correctly converted to ActNN layers.
print(model)    # Should be actnn.QConv2d, actnn.QBatchNorm2d, etc.

Advanced Features

  • Convert the model manually.
    ActNN is implemented as a collection of memory-saving layers, including actnn.QConv1d, QConv2d, QConv3d, QConvTranspose1d, QConvTranspose2d, QConvTranspose3d, QBatchNorm1d, QBatchNorm2d, QBatchNorm3d, QLinear, QReLU, QSyncBatchNorm, QMaxPool2d. These layers have identical interface to their PyTorch counterparts. You can construct the model manually using these layers as the building blocks. See ResNetBuilder and resnet_configs in image_classification/image_classification/resnet.py for example.
  • (Optional) Change the data loader
    If you want to use per-sample gradient information for adaptive quantization, you have to update the dataloader to return sample indices. You can see train_loader in mem_speed_benchmark/train.py for example. In addition, you have to update the configurations.
from actnn import config, QScheme
config.use_gradient = True
QScheme.num_samples = 1300000   # the size of training set

You can find sample code in the above script.

Examples

Benchmark Memory Usage and Training Speed

See mem_speed_benchmark. Please do NOT measure the memory usage by nvidia-smi. nvidia-smi reports the size of the memory pool allocated by PyTorch, which can be much larger than the size of acutal used memory.

Image Classification

See image_classification

Object Detection, Semantic Segmentation, Self-Supervised Learning, ...

Here is the example memory-efficient training for ResNet50, built upon the OpenMMLab toolkits. We use ActNN with the default optimization level (L3). Our training runs are available at Weights & Biases.

Installation

  1. Install mmcv
export MMCV_ROOT=/path/to/clone/actnn-mmcv
git clone https://github.com/DequanWang/actnn-mmcv $MMCV_ROOT
cd $MMCV_ROOT
MMCV_WITH_OPS=1 MMCV_WITH_ORT=0 pip install -e .
  1. Install mmdet, mmseg, mmssl, ...
export MMDET_ROOT=/path/to/clone/actnn-mmdet
git clone https://github.com/DequanWang/actnn-mmdet $MMDET_ROOT
cd $MMDET_ROOT
python setup.py develop
export MMSEG_ROOT=/path/to/clone/actnn-mmseg
git clone https://github.com/DequanWang/actnn-mmseg $MMSEG_ROOT
cd $MMSEG_ROOT
python setup.py develop
export MMSSL_ROOT=/path/to/clone/actnn-mmssl
git clone https://github.com/DequanWang/actnn-mmssl $MMSSL_ROOT
cd $MMSSL_ROOT
python setup.py develop

Single GPU training

cd $MMDET_ROOT
python tools/train.py configs/actnn/faster_rcnn_r50_fpn_1x_coco_1gpu.py
# https://wandb.ai/actnn/detection/runs/ye0aax5s
# ActNN mAP 37.4 vs Official mAP 37.4
python tools/train.py configs/actnn/retinanet_r50_fpn_1x_coco_1gpu.py
# https://wandb.ai/actnn/detection/runs/1x9cwokw
# ActNN mAP 36.3 vs Official mAP 36.5
cd $MMSEG_ROOT
python tools/train.py configs/actnn/fcn_r50-d8_512x1024_80k_cityscapes_1gpu.py
# https://wandb.ai/actnn/segmentation/runs/159if8da
# ActNN mIoU 72.9 vs Official mIoU 73.6
python tools/train.py configs/actnn/fpn_r50_512x1024_80k_cityscapes_1gpu.py
# https://wandb.ai/actnn/segmentation/runs/25j9iyv3
# ActNN mIoU 74.7 vs Official mIoU 74.5

Multiple GPUs training

cd $MMSSL_ROOT
bash tools/dist_train.sh configs/selfsup/actnn/moco_r50_v2_bs512_e200_imagenet_2gpu.py 2
# https://wandb.ai/actnn/mmssl/runs/lokf7ydo
# https://wandb.ai/actnn/mmssl/runs/2efmbuww
# ActNN top1 67.3 vs Official top1 67.7

For more detailed guidance, please refer to the docs of mmcv, mmdet, mmseg, mmssl.

FAQ

  1. Does ActNN supports CPU training?
    Currently, ActNN only supports CUDA.

  2. Accuracy degradation / diverged training with ActNN.
    ActNN applies lossy compression to the activations. In some challenging cases, our default compression strategy might be too aggressive. In this case, you may try more conservative compression strategies (which consume more memory):

    • 4-bit per-group quantization
    actnn.set_optimization_level("L2")
    • 8-bit per-group quantization
    actnn.set_optimization_level("L2")
    actnn.config.activation_compression_bits = [8]

    If none of these works, you may report to us by creating an issue.

Correspondence

Please email Jianfei Chen and Lianmin Zheng. Any questions or discussions are welcomed!

Citation

If the actnn library is helpful in your research, please consider citing our paper:

@article{chen2021actnn,
  title={ActNN: Reducing Training Memory Footprint via 2-Bit Activation Compressed Training},
  author={Chen, Jianfei and Zheng, Lianmin and Yao, Zhewei and Wang, Dequan and Stoica, Ion and Mahoney, Michael W and Gonzalez, Joseph E},
  journal={arXiv preprint arXiv:2104.14129},
  year={2021}
}
Owner
UC Berkeley RISE
REAL-TIME INTELLIGENT SECURE EXPLAINABLE SYSTEMS
UC Berkeley RISE
Cross-Task Consistency Learning Framework for Multi-Task Learning

Cross-Task Consistency Learning Framework for Multi-Task Learning Tested on numpy(v1.19.1) opencv-python(v4.4.0.42) torch(v1.7.0) torchvision(v0.8.0)

Aki Nakano 2 Jan 08, 2022
MogFace: Towards a Deeper Appreciation on Face Detection

MogFace: Towards a Deeper Appreciation on Face Detection Introduction In this repo, we propose a promising face detector, termed as MogFace. Our MogFa

48 Dec 20, 2022
Code for Estimating Multi-cause Treatment Effects via Single-cause Perturbation (NeurIPS 2021)

Estimating Multi-cause Treatment Effects via Single-cause Perturbation (NeurIPS 2021) Single-cause Perturbation (SCP) is a framework to estimate the m

Zhaozhi Qian 9 Sep 28, 2022
Reproducible research and reusable acyclic workflows in Python. Execute code on HPC systems as if you executed them on your personal computer!

Reproducible research and reusable acyclic workflows in Python. Execute code on HPC systems as if you executed them on your machine! Motivation Would

Joeri Hermans 15 Sep 11, 2022
基于深度强化学习的原神自动钓鱼AI

原神自动钓鱼AI由YOLOX, DQN两部分模型组成。使用迁移学习,半监督学习进行训练。 模型也包含一些使用opencv等传统数字图像处理方法实现的不可学习部分。

4.2k Jan 01, 2023
Official Pytorch implementation of 6DRepNet: 6D Rotation representation for unconstrained head pose estimation.

6D Rotation Representation for Unconstrained Head Pose Estimation (Pytorch) Paper Thorsten Hempel and Ahmed A. Abdelrahman and Ayoub Al-Hamadi, "6D Ro

Thorsten Hempel 284 Dec 23, 2022
CAMoE + Dual SoftMax Loss (DSL): Improving Video-Text Retrieval by Multi-Stream Corpus Alignment and Dual Softmax Loss

CAMoE + Dual SoftMax Loss (DSL): Improving Video-Text Retrieval by Multi-Stream Corpus Alignment and Dual Softmax Loss This is official implement of "

程星 87 Dec 24, 2022
Official PyTorch implementation of GDWCT (CVPR 2019, oral)

This repository provides the official code of GDWCT, and it is written in PyTorch. Paper Image-to-Image Translation via Group-wise Deep Whitening-and-

WonwoongCho 135 Dec 02, 2022
Official implementation of particle-based models (GNS and DPI-Net) on the Physion dataset.

Physion: Evaluating Physical Prediction from Vision in Humans and Machines [paper] Daniel M. Bear, Elias Wang, Damian Mrowca, Felix J. Binder, Hsiao-Y

Hsiao-Yu Fish Tung 18 Dec 19, 2022
An index of recommendation algorithms that are based on Graph Neural Networks.

An index of recommendation algorithms that are based on Graph Neural Networks.

FIB LAB, Tsinghua University 564 Jan 07, 2023
Retrieve and analysis data from SDSS (Sloan Digital Sky Survey)

Author: Behrouz Safari License: MIT sdss A python package for retrieving and analysing data from SDSS (Sloan Digital Sky Survey) Installation Install

Behrouz 3 Oct 28, 2022
2021:"Bridging Global Context Interactions for High-Fidelity Image Completion"

TFill arXiv | Project This repository implements the training, testing and editing tools for "Bridging Global Context Interactions for High-Fidelity I

Chuanxia Zheng 111 Jan 08, 2023
A spatial genome aligner for analyzing multiplexed DNA-FISH imaging data.

jie jie is a spatial genome aligner. This package parses true chromatin imaging signal from noise by aligning signals to a reference DNA polymer model

Bojing Jia 9 Sep 29, 2022
Code for our TKDE paper "Understanding WeChat User Preferences and “Wow” Diffusion"

wechat-wow-analysis Understanding WeChat User Preferences and “Wow” Diffusion. Fanjin Zhang, Jie Tang, Xueyi Liu, Zhenyu Hou, Yuxiao Dong, Jing Zhang,

18 Sep 16, 2022
Watch faces morph into each other with StyleGAN 2, StyleGAN, and DCGAN!

FaceMorpher FaceMorpher is an innovative project to get a unique face morph (or interpolation for geeks) on a website. Yes, this means you can see fac

Anish 9 Jun 24, 2022
Python calculations for the position of the sun and moon.

Astral This is 'astral' a Python module which calculates Times for various positions of the sun: dawn, sunrise, solar noon, sunset, dusk, solar elevat

Simon Kennedy 169 Dec 20, 2022
Towards Implicit Text-Guided 3D Shape Generation (CVPR2022)

Towards Implicit Text-Guided 3D Shape Generation Towards Implicit Text-Guided 3D Shape Generation (CVPR2022) Code for the paper [Towards Implicit Text

55 Dec 16, 2022
This implementation contains the application of GPlearn's symbolic transformer on a commodity futures sector of the financial market.

GPlearn_finiance_stock_futures_extension This implementation contains the application of GPlearn's symbolic transformer on a commodity futures sector

Chengwei <a href=[email protected]"> 189 Dec 25, 2022
OpenPCDet Toolbox for LiDAR-based 3D Object Detection.

OpenPCDet OpenPCDet is a clear, simple, self-contained open source project for LiDAR-based 3D object detection. It is also the official code release o

OpenMMLab 3.2k Dec 31, 2022
This is an official pytorch implementation of Lite-HRNet: A Lightweight High-Resolution Network.

Lite-HRNet: A Lightweight High-Resolution Network Introduction This is an official pytorch implementation of Lite-HRNet: A Lightweight High-Resolution

HRNet 675 Dec 25, 2022