Implementation of the Triangle Multiplicative module, used in Alphafold2 as an efficient way to mix rows or columns of a 2d feature map, as a standalone package for Pytorch

Last update: Oct 28, 2022

Overview

Triangle Multiplicative Module - Pytorch

Implementation of the Triangle Multiplicative module, used in Alphafold2 as an efficient way to mix rows or columns of a 2d feature map, as a standalone package for Pytorch

Install

$ pip install triangle-multiplicative-module

Usage

import torch
from triangle_multiplicative_module import TriangleMultiplicativeModule

model = TriangleMultiplicativeModule(
    dim = 64,            # feature map dimension
    hidden_dim = 128,    # intermediate dimension size
    mix = 'outgoing'     # either 'ingoing' or 'outgoing'
)

fmap = torch.randn(1, 256, 256, 64)
mask = torch.ones(1, 256, 256).bool()

model(fmap, mask = mask) # (1, 256, 256, 64)

Citations

@Article{AlphaFold2021,
    author  = {Jumper, John and Evans, Richard and Pritzel, Alexander and Green, Tim and Figurnov, Michael and Ronneberger, Olaf and Tunyasuvunakool, Kathryn and Bates, Russ and {\v{Z}}{\'\i}dek, Augustin and Potapenko, Anna and Bridgland, Alex and Meyer, Clemens and Kohl, Simon A A and Ballard, Andrew J and Cowie, Andrew and Romera-Paredes, Bernardino and Nikolov, Stanislav and Jain, Rishub and Adler, Jonas and Back, Trevor and Petersen, Stig and Reiman, David and Clancy, Ellen and Zielinski, Michal and Steinegger, Martin and Pacholska, Michalina and Berghammer, Tamas and Bodenstein, Sebastian and Silver, David and Vinyals, Oriol and Senior, Andrew W and Kavukcuoglu, Koray and Kohli, Pushmeet and Hassabis, Demis},
    journal = {Nature},
    title   = {Highly accurate protein structure prediction with {AlphaFold}},
    year    = {2021},
    doi     = {10.1038/s41586-021-03819-2},
    note    = {(Accelerated article preview)},
}

Comments

Definition about 'ingoing' and 'outgoing'

Hi Luci!

I've seen the lines define the 'ingoing' and 'outgoing': if mix == 'ingoing': self.mix_einsum_eq = '... i k d, ... j k d -> ... i j d' elif mix == 'outgoing': self.mix_einsum_eq = '... k j d, ... k i d -> ... i j d'

but based on the original paper, I found maybe should be the inverse case, that: if mix == 'ingoing': self.mix_einsum_eq = '... k i d, ... k j d -> ... i j d' elif mix == 'outgoing': self.mix_einsum_eq = '... j k d, ... i k d -> ... i j d'

based on what I understand, 'ingoing' should be the case that information comes from other edges (like 'ki','kj') aggregate togther and bind with edge 'ij'.

Thanks for your contribution!

opened by CiaoHe 1

Official implementation of "StyleCariGAN: Caricature Generation via StyleGAN Feature Map Modulation" (SIGGRAPH 2021)

StyleCariGAN: Caricature Generation via StyleGAN Feature Map Modulation This repository contains the official PyTorch implementation of the following

270 Dec 30, 2022

Codes for the paper Contrast and Mix: Temporal Contrastive Video Domain Adaptation with Background Mixing

Contrast and Mix (CoMix) The repository contains the codes for the paper Contrast and Mix: Temporal Contrastive Video Domain Adaptation with Backgroun

Computer Vision and Intelligence Research (CVIR)

13 Dec 10, 2022

Lightweight mmm - Lightweight (Bayesian) Media Mix Model

Lightweight (Bayesian) Media Mix Model This is not an official Google product. L

342 Jan 3, 2023

RetinaNet-PyTorch - A RetinaNet Pytorch Implementation on remote sensing images and has the similar mAP result with RetinaNet in MMdetection

🚀 RetinaNet Horizontal Detector Based PyTorch This is a horizontal detector Ret

13 Nov 19, 2022

A python software that can help blind people find things like laptops, phones, etc the same way a guide dog guides a blind person in finding his way.

GuidEye A python software that can help blind people find things like laptops, phones, etc the same way a guide dog guides a blind person in finding h

0 Aug 9, 2022

Implementation of Common Image Evaluation Metrics by Sayed Nadim (sayednadim.github.io). The repo is built based on full reference image quality metrics such as L1, L2, PSNR, SSIM, LPIPS. and feature-level quality metrics such as FID, IS. It can be used for evaluating image denoising, colorization, inpainting, deraining, dehazing etc. where we have access to ground truth.

Image Quality Evaluation Metrics Implementation of some common full reference image quality metrics. The repo is built based on full reference image q

10 Jan 1, 2023

Implementation of the Triangle Multiplicative module, used in Alphafold2 as an efficient way to mix rows or columns of a 2d feature map, as a standalone package for Pytorch

Related tags

Overview

Triangle Multiplicative Module - Pytorch

Install

Usage

Citations

You might also like...

Official implementation of "StyleCariGAN: Caricature Generation via StyleGAN Feature Map Modulation" (SIGGRAPH 2021)

Codes for the paper Contrast and Mix: Temporal Contrastive Video Domain Adaptation with Background Mixing

Lightweight mmm - Lightweight (Bayesian) Media Mix Model

RetinaNet-PyTorch - A RetinaNet Pytorch Implementation on remote sensing images and has the similar mAP result with RetinaNet in MMdetection

A python software that can help blind people find things like laptops, phones, etc the same way a guide dog guides a blind person in finding his way.

Standalone pre-training recipe with JAX+Flax

PyTorch implementation of saliency map-aided GAN for Auto-demosaic+denosing

Efficient-GlobalPointer - Pytorch Efficient GlobalPointer

Comments

Definition about 'ingoing' and 'outgoing'

Releases(0.0.3)

0.0.3(Aug 3, 2021)

0.0.2(Jul 23, 2021)

0.0.1(Jul 19, 2021)

Owner

Phil Wang

Auto HMM: Automatic Discrete and Continous HMM including Model selection

Aiming at the common training datsets split, spectrum preprocessing, wavelength select and calibration models algorithm involved in the spectral analysis process

The official PyTorch implementation of Curriculum by Smoothing (NeurIPS 2020, Spotlight).

Code and data of the ACL 2021 paper: Few-Shot Text Ranking with Meta Adapted Synthetic Weak Supervision

Proximal Backpropagation - a neural network training algorithm that takes implicit instead of explicit gradient steps

Machine Translation Implement By Bi-GRU And Transformer

PyExplainer: A Local Rule-Based Model-Agnostic Technique (Explainable AI)

Neural Fixed-Point Acceleration for Convex Optimization

Baleen: Robust Multi-Hop Reasoning at Scale via Condensed Retrieval (NeurIPS'21)

This initial strategy was developed specifically for larger pools and is based on taking a moving average and deriving Bollinger Bands to create a projected active liquidity range.

Artifacts for paper "MMO: Meta Multi-Objectivization for Software Configuration Tuning"

Official implementation of "GS-WGAN: A Gradient-Sanitized Approach for Learning Differentially Private Generators" (NeurIPS 2020)

Code for the CVPR2021 paper "Patch-NetVLAD: Multi-Scale Fusion of Locally-Global Descriptors for Place Recognition"

Codes accompanying the paper "Believe What You See: Implicit Constraint Approach for Offline Multi-Agent Reinforcement Learning" (NeurIPS 2021 Spotlight

Code for "Multi-View Multi-Person 3D Pose Estimation with Plane Sweep Stereo"

A Robust Unsupervised Ensemble of Feature-Based Explanations using Restricted Boltzmann Machines

Source code for our Paper "Learning in High-Dimensional Feature Spaces Using ANOVA-Based Matrix-Vector Multiplication"

The Pytorch implementation for "Video-Text Pre-training with Learned Regions"

Deep learning operations reinvented (for pytorch, tensorflow, jax and others)

3D-Reconstruction 基于深度学习方法的单目多视图三维重建