This repo is the official implementation of "L2ight: Enabling On-Chip Learning for Optical Neural Networks via Efficient in-situ Subspace Optimization".

Related tags

Deep LearningL2ight
Overview

L2ight

By Jiaqi Gu, Hanqing Zhu, Chenghao Feng, Zixuan Jiang, Ray T. Chen and David Z. Pan.

This repo is the official implementation of "L2ight: Enabling On-Chip Learning for Optical Neural Networks via Efficient in-situ Subspace Optimization".

Introduction

L2ight is a closed-loop ONN on-chip learning framework to enable scalable ONN mapping and efficient in-situ learning. L2ight adopts a three-stage learning flow that first calibrates the complicated photonic circuit states under challenging physical constraints, then performs photonic core mapping via combined analytical solving and zeroth-order optimization. A subspace learning procedure with multi-level sparsity is integrated into L2ight to enable in-situ gradient evaluation and fast adaptation, unleashing the power of optics for real on-chip intelligence. L2ight outperforms prior ONN training protocols with 3-order-of-magnitude higher scalability and over 30X better efficiency, when benchmarked on various models and learning tasks. This synergistic framework is the first scalable on-chip learning solution that pushes this emerging field from intractable to scalable and further to efficient for next-generation self-learnable photonic neural chips.

flow teaser

Dependencies

  • Python >= 3.6
  • pyutils >= 0.0.1. See pyutils for installation.
  • pytorch-onn >= 0.0.1. See pytorch-onn for installation.
  • Python libraries listed in requirements.txt
  • NVIDIA GPUs and CUDA >= 10.2

Structures

  • core/
    • models/
      • layers/
        • custom_conv2d and custom_linear layers
        • utils.py: sampler and profiler
      • sparse_bp_*.py: model definition
      • sparse_bp_base.py: base model definition; identity calibration and mapping codes.
    • optimizer/: mixedtrain and flops optimizers
    • builder.py: build training utilities
  • script/: contains experiment scripts
  • train_pretrain.py, train_map.py, train_learn.py, train_zo_learn.py: training logic
  • compare_gradient.py: compare approximated gradients with true gradients for ablation

Usage

  • Pretrain model.
    > python3 train_pretrain.py config/cifar10/vgg8/pretrain.yml

  • Identity calibration and parallel mapping. Please set your hyperparameters in CONFIG=config/cifar10/vgg8/pm/pm.yml and run
    > python3 train_map.py CONFIG --checkpoint.restore_checkpoint=path/to/your/pretrained/checkpoint

  • Subspace learning with multi-level sampling. Please set your hyperparameters in CONFIG=config/cifar10/vgg8/ds/learn.yml and run
    > python3 train_learn.py CONFIG --checkpoint.restore_chekcpoint=path/to/your/mapped/checkpoint --checkpoint.resume=1

  • All scripts for experiments are in ./script. For example, to run subspace learning with feedback sampling, column sampling, and data sampling, you can write proper task setting in SCRIPT=script/vgg8/train_ds_script.py and run
    > python3 SCRIPT

  • Comparison experiments with RAD [ICLR 2021] and SWAT-U [NeurIPS 2020]. Run with the SCRIPT=script/vgg8/train_rad_script.py and script/vgg8/train_swat_script.py,
    > python3 SCRIPT

  • Comparison with FLOPS [DAC 2020] and MixedTrn [AAAI 2021]. Run with the METHOD=mixedtrain or flops,
    > python3 train_zo_learn.py config/mnist/cnn3/METHOD/learn.yml

Citing L2ight

@inproceedings{gu2021L2ight,
  title={L2ight: Enabling On-Chip Learning for Optical Neural Networks via Efficient in-situ Subspace Optimization},
  author={Jiaqi Gu and Hanqing Zhu and Chenghao Feng and Zixuan Jiang and Ray T. Chen and David Z. Pan},
  journal={Conference on Neural Information Processing Systems (NeurIPS)},
  year={2021}
}
Owner
Jiaqi Gu
PhD Student at UT Austin
Jiaqi Gu
A highly efficient, fast, powerful and light-weight anime downloader and streamer for your favorite anime.

AnimDL - Download & Stream Your Favorite Anime AnimDL is an incredibly powerful tool for downloading and streaming anime. Core features Abuses the dev

KR 759 Jan 08, 2023
YOLOv7 - Framework Beyond Detection

🔥🔥🔥🔥 YOLO with Transformers and Instance Segmentation, with TensorRT acceleration! 🔥🔥🔥

JinTian 3k Jan 01, 2023
《Single Image Reflection Removal Beyond Linearity》(CVPR 2019)

Single-Image-Reflection-Removal-Beyond-Linearity Paper Single Image Reflection Removal Beyond Linearity. Qiang Wen, Yinjie Tan, Jing Qin, Wenxi Liu, G

Qiang Wen 51 Jun 24, 2022
ELECTRA: Pre-training Text Encoders as Discriminators Rather Than Generators

ELECTRA Introduction ELECTRA is a method for self-supervised language representation learning. It can be used to pre-train transformer networks using

Google Research 2.1k Dec 28, 2022
code for `Look Closer to Segment Better: Boundary Patch Refinement for Instance Segmentation`

Look Closer to Segment Better: Boundary Patch Refinement for Instance Segmentation (CVPR 2021) Introduction PBR is a conceptually simple yet effective

H.Chen 143 Jan 05, 2023
Efficient neural networks for analog audio effect modeling

micro-TCN Efficient neural networks for audio effect modeling

Christian Steinmetz 94 Dec 29, 2022
You are AllSet: A Multiset Function Framework for Hypergraph Neural Networks.

AllSet This is the repo for our paper: You are AllSet: A Multiset Function Framework for Hypergraph Neural Networks. We prepared all codes and a subse

Jianhao 51 Dec 24, 2022
Official git for "CTAB-GAN: Effective Table Data Synthesizing"

CTAB-GAN This is the official git paper CTAB-GAN: Effective Table Data Synthesizing. The paper is published on Asian Conference on Machine Learning (A

30 Dec 26, 2022
CLIP+FFT text-to-image

Aphantasia This is a text-to-image tool, part of the artwork of the same name. Based on CLIP model, with FFT parameterizer from Lucent library as a ge

vadim epstein 690 Jan 02, 2023
[CVPR2021] DoDNet: Learning to segment multi-organ and tumors from multiple partially labeled datasets

DoDNet This repo holds the pytorch implementation of DoDNet: DoDNet: Learning to segment multi-organ and tumors from multiple partially labeled datase

116 Dec 12, 2022
A PyTorch Lightning solution to training OpenAI's CLIP from scratch.

train-CLIP 📎 A PyTorch Lightning solution to training CLIP from scratch. Goal ⚽ Our aim is to create an easy to use Lightning implementation of OpenA

Cade Gordon 396 Dec 30, 2022
Neural Factorization of Shape and Reflectance Under An Unknown Illumination

NeRFactor [Paper] [Video] [Project] This is the authors' code release for: NeRFactor: Neural Factorization of Shape and Reflectance Under an Unknown I

Google 283 Jan 04, 2023
Replication Package for AequeVox:Automated Fariness Testing for Speech Recognition Systems

AequeVox Replication Package for AequeVox:Automated Fariness Testing for Speech Recognition Systems README under development. Python Packages Required

Sai Sathiesh 2 Aug 28, 2022
TSDF++: A Multi-Object Formulation for Dynamic Object Tracking and Reconstruction

TSDF++: A Multi-Object Formulation for Dynamic Object Tracking and Reconstruction TSDF++ is a novel multi-object TSDF formulation that can encode mult

ETHZ ASL 130 Dec 29, 2022
ICCV2021 Oral SA-ConvONet: Sign-Agnostic Optimization of Convolutional Occupancy Networks

Sign-Agnostic Convolutional Occupancy Networks Paper | Supplementary | Video | Teaser Video | Project Page This repository contains the implementation

64 Jan 05, 2023
ProjectOxford-ClientSDK - This repo has moved :house: Visit our website for the latest SDKs & Samples

This project has moved 🏠 We heard your feedback! This repo has been deprecated and each project has moved to a new home in a repo scoped by API and p

Microsoft 970 Nov 28, 2022
This framework implements the data poisoning method found in the paper Adversarial Examples Make Strong Poisons

Adversarial poison generation and evaluation. This framework implements the data poisoning method found in the paper Adversarial Examples Make Strong

31 Nov 01, 2022
Python project to take sound as input and output as RGB + Brightness values suitable for DMX

sound-to-light Python project to take sound as input and output as RGB + Brightness values suitable for DMX Current goals: Get one pixel working: Vary

Bobby Cox 1 Nov 17, 2021
MicroNet: Improving Image Recognition with Extremely Low FLOPs (ICCV 2021)

MicroNet: Improving Image Recognition with Extremely Low FLOPs (ICCV 2021) A pytorch implementation of MicroNet. If you use this code in your research

Yunsheng Li 293 Dec 28, 2022
PyTorch META-DATASET (Few-shot classification benchmark)

PyTorch META-DATASET (Few-shot classification benchmark) This repo contains a PyTorch implementation of meta-dataset and a unified implementation of s

Malik Boudiaf 39 Oct 31, 2022