Implements MLP-Mixer: An all-MLP Architecture for Vision.

Last update: Jan 04, 2023

Overview

MLP-Mixer-CIFAR10

This repository implements MLP-Mixer as proposed in MLP-Mixer: An all-MLP Architecture for Vision. The paper introduces an all MLP (Multi-layer Perceptron) architecture for computer vision tasks. Yannic Kilcher walks through the architecture in this video.

Experiments reported in this repository are on CIFAR-10.

What's included?

Distributed training with mixed-precision.
Visualization of the token-mixing MLP weights.
A TensorBoard callback to keep track of the learned linear projections of the image patches.

Screen.Recording.2021-05-25.at.5.49.20.PM.mov

Notebooks

MLP_Mixer_Training.ipynb: MLP-Mixer utilities along with model training.
ResNet20.ipynb: Trains a ResNet20 for comparison purposes.
Visualization.ipynb: Visualizes the learned projections and token-mixing MLPs.

Note: These notebooks are runnable on Colab. If you don't have access to a tensor-core GPU, please disable the mixed-precision block while running the code.

Results

MLP-Mixer achieves competitive results. The figure below summarizes top-1 accuracies on CIFAR-10 test set with respect to varying MLP blocks.

Notable hyperparameters are:

Image size: 72x72
Patch size: 9x9
Hidden dimension for patches: 64
Hidden dimension for patches: 128

The table below reports the parameter counts for the different MLP-Mixer variants:

ResNet20 (0.571969 Million) achieves 78.14% under the exact same training configuration. Refer to this notebook for more details.

Models

You can reproduce the results reported above. The model files are available here.

Acknowledgements

ML-GDE Program for providing GCP credits.

You might also like...

An All-MLP solution for Vision, from Google AI

MLP Mixer - Pytorch An All-MLP solution for Vision, from Google AI, in Pytorch. No convolutions nor attention needed! Yannic Kilcher video Install $ p

784 Jan 6, 2023

Implementation of "A MLP-like Architecture for Dense Prediction"

A MLP-like Architecture for Dense Prediction (arXiv) Updates (22/07/2021) Initial release. Model Zoo We provide CycleMLP models pretrained on ImageNet

244 Dec 27, 2022

Model search is a framework that implements AutoML algorithms for model architecture search at scale

Model search (MS) is a framework that implements AutoML algorithms for model architecture search at scale. It aims to help researchers speed up their exploration process for finding the right model architecture for their classification problems (i.e., DNNs with different types of layers).

3.2k Dec 31, 2022

A task-agnostic vision-language architecture as a step towards General Purpose Vision

Towards General Purpose Vision Systems By Tanmay Gupta, Amita Kamath, Aniruddha Kembhavi, and Derek Hoiem Overview Welcome to the official code base f

79 Dec 23, 2022

MLP-Like Vision Permutator for Visual Recognition (PyTorch)

Vision Permutator: A Permutable MLP-Like Architecture for Visual Recognition (arxiv) This is a Pytorch implementation of our paper. We present Vision

162 Nov 28, 2022

code for paper "Does Unsupervised Architecture Representation Learning Help Neural Architecture Search?"

Does Unsupervised Architecture Representation Learning Help Neural Architecture Search? Code for paper: Does Unsupervised Architecture Representation

39 Dec 17, 2022

Implementation of ResMLP, an all MLP solution to image classification, in Pytorch

ResMLP - Pytorch Implementation of ResMLP, an all MLP solution to image classification out of Facebook AI, in Pytorch Install $ pip install res-mlp-py

178 Dec 2, 2022

Implementation of gMLP, an all-MLP replacement for Transformers, in Pytorch

383 Jan 2, 2023

Episodic Transformer (E.T.) is a novel attention-based architecture for vision-and-language navigation. E.T. is based on a multimodal transformer that encodes language inputs and the full episode history of visual observations and actions.

Episodic Transformers (E.T.) Episodic Transformer for Vision-and-Language Navigation Alexander Pashevich, Cordelia Schmid, Chen Sun Episodic Transform

62 Dec 24, 2022

Comments

Could patches number != MLP token mixing dimension?

I try to change the model into B/16 MLP-Mixer. is this setting, the patch number ( sequence length) != MLP token mixing dimension. But the code will report an error when it implements "x = layers.Add()([x, token_mixing])" because the two operation numbers have different shapes. Take an example, B/16 Settings: image 3232, 2D hidden layer 768, PP= 16*16, token mixing mlp dimentsion= 384, channel mlp dimension = 3072. Thus patch number ( sequence length) = 4, table value shape= (4, 768) When the code runs x = layers.Add()([x, token_mixing]) in the token mixing layer. rx shape=[4, 768], token_mixing shape = [384, 768]

It is strange why the MLP-Mixer paper could set different parameters "patch number ( sequence length) != MLP token mixing dimensio"

opened by LouiValley 2
Why the accuracy drops after epoch 100/100 (accuracy drops from 91% to 71%)

I trained the Network ( NUM_MIXER_LAYERS =4 )

At epoch 100:

Epoch 100/100

1/44 [..............................] - ETA: 1s - loss: 0.2472 - accuracy: 0.9160 3/44 [=>............................] - ETA: 1s - loss: 0.2424 - accuracy: 0.9162 5/44 [==>...........................] - ETA: 1s - loss: 0.2431 - accuracy: 0.9155 7/44 [===>..........................] - ETA: 1s - loss: 0.2424 - accuracy: 0.9154 9/44 [=====>........................] - ETA: 1s - loss: 0.2419 - accuracy: 0.9155 11/44 [======>.......................] - ETA: 1s - loss: 0.2423 - accuracy: 0.9150 13/44 [=======>......................] - ETA: 1s - loss: 0.2426 - accuracy: 0.9145 15/44 [=========>....................] - ETA: 1s - loss: 0.2430 - accuracy: 0.9142 17/44 [==========>...................] - ETA: 1s - loss: 0.2433 - accuracy: 0.9140 19/44 [===========>..................] - ETA: 1s - loss: 0.2435 - accuracy: 0.9138 21/44 [=============>................] - ETA: 0s - loss: 0.2438 - accuracy: 0.9136 23/44 [==============>...............] - ETA: 0s - loss: 0.2439 - accuracy: 0.9135 25/44 [================>.............] - ETA: 0s - loss: 0.2440 - accuracy: 0.9134 27/44 [=================>............] - ETA: 0s - loss: 0.2440 - accuracy: 0.9133 29/44 [==================>...........] - ETA: 0s - loss: 0.2442 - accuracy: 0.9132 31/44 [====================>.........] - ETA: 0s - loss: 0.2445 - accuracy: 0.9130 33/44 [=====================>........] - ETA: 0s - loss: 0.2447 - accuracy: 0.9129 35/44 [======================>.......] - ETA: 0s - loss: 0.2450 - accuracy: 0.9127 37/44 [========================>.....] - ETA: 0s - loss: 0.2454 - accuracy: 0.9125 39/44 [=========================>....] - ETA: 0s - loss: 0.2459 - accuracy: 0.9123 41/44 [==========================>...] - ETA: 0s - loss: 0.2463 - accuracy: 0.9121 43/44 [============================>.] - ETA: 0s - loss: 0.2469 - accuracy: 0.9119 44/44 [==============================] - 2s 46ms/step - loss: 0.2474 - accuracy: 0.9117 - val_loss: 1.1145 - val_accuracy: 0.7226

Then it still have an extra training, 1/313 [..............................] - ETA: 24:32 - loss: 0.5860 - accuracy: 0.8125 8/313 [..............................] - ETA: 2s - loss: 1.2071 - accuracy: 0.6953 ..... 313/313 [==============================] - ETA: 0s - loss: 1.0934 - accuracy: 0.7161 313/313 [==============================] - 12s 22ms/step - loss: 1.0934 - accuracy: 0.7161 Test accuracy: 71.61

opened by LouiValley 1
Consider either turning off auto-sharding or switching the auto_shard_policy to DATA

Excuse me, when I try to run it on the serve, it tips:

Consider either turning off auto-sharding or switching the auto_shard_policy to DATA to shard this dataset. You can do this by creating a new tf.data.Options() object then setting options.experimental_distribute.auto_shard_policy = AutoShardPolicy.DATA before applying the options object to the dataset via dataset.with_options(options). 2021-11-21 11:59:20.861052: W tensorflow/python/util/util.cc:348] Sets are not currently considered sequences, but this may change in the future, so consider avoiding using them.

BTW, my TensorFlow version is 2.4.0, how to fix this problem?

opened by LouiValley 1

Releases(Models)

Models(May 25, 2021)
Models included in the attached archive are:

MLP-Mixer with varying number of MLP blocks (2, 4, 6, 8, and 10)

ResNet20

Logs include TensorBoard logs to track how the linear projections evolve over time.
Source code(tar.gz)
Source code(zip)
logs.tar.gz(4.42 MB)
models.zip(22.15 MB)

Owner

Sayak Paul

Trying to learn how machines learn.

GitHub Repository

Official PyTorch implementation of the paper Image-Based CLIP-Guided Essence Transfer.

TargetCLIP- official pytorch implementation of the paper Image-Based CLIP-Guided Essence Transfer This repository finds a global direction in StyleGAN

221 Dec 13, 2022

Official PyTorch implementation of "Uncertainty-Based Offline Reinforcement Learning with Diversified Q-Ensemble" (NeurIPS'21)

Uncertainty-Based Offline Reinforcement Learning with Diversified Q-Ensemble This is the code for reproducing the results of the paper Uncertainty-Bas

43 Nov 23, 2022

Asymmetric Bilateral Motion Estimation for Video Frame Interpolation, ICCV2021

ABME (ICCV2021) Junheum Park, Chul Lee, and Chang-Su Kim Official PyTorch Code for "Asymmetric Bilateral Motion Estimation for Video Frame Interpolati

86 Dec 28, 2022

Neural network chess engine trained on Gary Kasparov's games.

Neural Chess It's not the best chess engine, but it is a chess engine. Proof of concept neural network chess engine (feed-forward multi-layer perceptr

3 Jun 22, 2022

This repository contains PyTorch models for SpecTr (Spectral Transformer).

SpecTr: Spectral Transformer for Hyperspectral Pathology Image Segmentation This repository contains PyTorch models for SpecTr (Spectral Transformer).

45 Dec 13, 2022

Machine Learning Time-Series Platform

cesium: Open-Source Platform for Time Series Inference Summary cesium is an open source library that allows users to: extract features from raw time s

632 Dec 26, 2022

PPLNN is a Primitive Library for Neural Network is a high-performance deep-learning inference engine for efficient AI inferencing

943 Jan 07, 2023

A 35mm camera, based on the Canonet G-III QL17 rangefinder, simulated in Python.

c is for Camera A 35mm camera, based on the Canonet G-III QL17 rangefinder, simulated in Python. The purpose of this project is to explore and underst

146 Sep 26, 2022

repro_eval is a collection of measures to evaluate the reproducibility/replicability of system-oriented IR experiments

repro_eval repro_eval is a collection of measures to evaluate the reproducibility/replicability of system-oriented IR experiments. The measures were d

9 May 25, 2022

FairEdit: Preserving Fairness in Graph Neural Networks through Greedy Graph Editing

FairEdit Relevent Publication FairEdit: Preserving Fairness in Graph Neural Networks through Greedy Graph Editing

5 Feb 04, 2022

Official Implementation of DAFormer: Improving Network Architectures and Training Strategies for Domain-Adaptive Semantic Segmentation

DAFormer: Improving Network Architectures and Training Strategies for Domain-Adaptive Semantic Segmentation [Arxiv] [Paper] As acquiring pixel-wise an

305 Dec 29, 2022