PyTorch implementation of Algorithm 1 of "On the Anatomy of MCMC-Based Maximum Likelihood Learning of Energy-Based Models"

Overview

Code for On the Anatomy of MCMC-Based Maximum Likelihood Learning of Energy-Based Models

This repository will reproduce the main results from our paper:

On the Anatomy of MCMC-Based Maximum Likelihood Learning of Energy-Based Models
Erik Nijkamp*, Mitch Hill*, Tian Han, Song-Chun Zhu, and Ying Nian Wu (*equal contributions)
https://arxiv.org/abs/1903.12370
AAAI 2020.

The files train_data.py and train_toy.py are PyTorch-based implementations of Algorithm 1 for image datasets and toy 2D distributions respectively. Both files will measure and plot the diagnostic values $d_{s_t}$ and $r_t$ described in Section 3 during training. The file eval.py will sample from a saved checkpoint using either unadjusted Langevin dynamics or Metropolis-Hastings adjusted Langevin dynamics. We provide an appendix ebm-anatomy-appendix.pdf that contains further practical considerations and empirical observations.

Config Files

The folder config_locker has several JSON files that reproduce different convergent and non-convergent learning outcomes for image datasets and toy distributions. Config files for evaluation of pre-trained networks are also included. The files data_config.json, toy_config.json, and eval_config.json fully explain the parameters for train_data.py, train_toy.py, and eval.py respectively.

Executable Files

To run an experiment with train_data.py, train_toy.py, or eval.py, just specify a name for the experiment folder and the location of the JSON config file:

# directory for experiment results
EXP_DIR = './name_of/new_folder/'
# json file with experiment config
CONFIG_FILE = './path_to/config.json'

before execution.

Other Files

Network structures are located in nets.py. A download function for Oxford Flowers 102 data, plotting functions, and a toy dataset class can be found in utils.py.

Diagnostics

Energy Difference and Langevin Gradient Magnitude: Both image and toy experiments will plot $d_{s_t}$ and $r_t$ (see Section 3) over training along with correlation plots as in Figure 4 (with ACF rather than PACF).

Landscape Plots: Toy experiments will plot the density and log-density (negative energy) for ground-truth, learned energy, and short-run models. Kernel density estimation is used to obtain the short-run density.

Short-Run MCMC Samples: Image data experiments will periodically visualize the short-run MCMC samples. A batch of persistent MCMC samples will also be saved for implementations that use persistent initialization for short-run sampling.

Long-Run MCMC Samples: Image data experiments have the option to obtain long-run MCMC samples during training. When log_longrun is set to true in a data config file, the training implementation will generate long-run MCMC samples at a frequency determined by log_longrun_freq. The appearance of long-run MCMC samples indicates whether the energy function assigns probability mass in realistic regions of the image space.

Pre-trained Networks

A convergent pre-trained network and non-convergent pre-trained network for the Oxford Flowers 102 dataset are available in the Releases section of the repository. The config files eval_flowers_convergent.json and eval_flowers_convergent_mh.json are set up to evaluate flowers_convergent_net.pth. The config file eval_flowers_nonconvergent.json is set up to evaluate flowers_nonconvergent_net.pth.

Contact

Please contact Mitch Hill ([email protected]) or Erik Nijkamp ([email protected]) for any questions.

You might also like...
Re-implementation of the Noise Contrastive Estimation algorithm for pyTorch, following "Noise-contrastive estimation: A new estimation principle for unnormalized statistical models." (Gutmann and Hyvarinen, AISTATS 2010)

Noise Contrastive Estimation for pyTorch Overview This repository contains a re-implementation of the Noise Contrastive Estimation algorithm, implemen

ppo_pytorch_cpp - an implementation of the proximal policy optimization algorithm for the C++ API of Pytorch
ppo_pytorch_cpp - an implementation of the proximal policy optimization algorithm for the C++ API of Pytorch

PPO Pytorch C++ This is an implementation of the proximal policy optimization algorithm for the C++ API of Pytorch. It uses a simple TestEnvironment t

PyTorch implementation of DreamerV2 model-based RL algorithm

PyDreamer Reimplementation of DreamerV2 model-based RL algorithm in PyTorch. The official DreamerV2 implementation can be found here. Features ... Run

PyTorch implementation of the implicit Q-learning algorithm (IQL)
PyTorch implementation of the implicit Q-learning algorithm (IQL)

Implicit-Q-Learning (IQL) PyTorch implementation of the implicit Q-learning algorithm IQL (Paper) Currently only implemented for online learning. Offl

PyTorch Implementation of the SuRP algorithm by the authors of the AISTATS 2022 paper "An Information-Theoretic Justification for Model Pruning"

PyTorch Implementation of the SuRP algorithm by the authors of the AISTATS 2022 paper "An Information-Theoretic Justification for Model Pruning".

A pytorch reprelication of the model-based reinforcement learning algorithm MBPO
A pytorch reprelication of the model-based reinforcement learning algorithm MBPO

Overview This is a re-implementation of the model-based RL algorithm MBPO in pytorch as described in the following paper: When to Trust Your Model: Mo

An algorithm that handles large-scale aerial photo co-registration, based on SURF, RANSAC and PyTorch autograd.
An algorithm that handles large-scale aerial photo co-registration, based on SURF, RANSAC and PyTorch autograd.

An algorithm that handles large-scale aerial photo co-registration, based on SURF, RANSAC and PyTorch autograd.

Implements pytorch code for the Accelerated SGD algorithm.

AccSGD This is the code associated with Accelerated SGD algorithm used in the paper On the insufficiency of existing momentum schemes for Stochastic O

PyGAD, a Python 3 library for building the genetic algorithm and training machine learning algorithms (Keras & PyTorch).
PyGAD, a Python 3 library for building the genetic algorithm and training machine learning algorithms (Keras & PyTorch).

PyGAD: Genetic Algorithm in Python PyGAD is an open-source easy-to-use Python 3 library for building the genetic algorithm and optimizing machine lear

Comments
  • Step size in Langevin Dynamics

    Step size in Langevin Dynamics

    Hi, in your code, when you do the langevin dynamics, you run x_s_t.data += - f_prime + config['epsilon'] * t.randn_like(x_s_t) However, does this mean that the step size for the gradient f_prim is 1? Should we run x_s_t.data += - 0.5*config['epsilon']**2*f_prime + config['epsilon'] * t.randn_like(x_s_t) instead?

    opened by XavierXiao 1
Releases(v1.0)
Owner
Mitch Hill
Assistant Professor of Statistics and Data Science at UCF
Mitch Hill
SoK: Vehicle Orientation Representations for Deep Rotation Estimation

SoK: Vehicle Orientation Representations for Deep Rotation Estimation Raymond H. Tu, Siyuan Peng, Valdimir Leung, Richard Gao, Jerry Lan This is the o

FIRE Capital One Machine Learning of the University of Maryland 12 Oct 07, 2022
This is the repo for Uncertainty Quantification 360 Toolkit.

UQ360 The Uncertainty Quantification 360 (UQ360) toolkit is an open-source Python package that provides a diverse set of algorithms to quantify uncert

International Business Machines 207 Dec 30, 2022
An MQA (Studio, originalSampleRate) identifier for lossless flac files written in Python.

An MQA (Studio, originalSampleRate) identifier for "lossless" flac files written in Python.

Daniel 10 Oct 03, 2022
Improving Object Detection by Estimating Bounding Box Quality Accurately

Improving Object Detection by Estimating Bounding Box Quality Accurately Abstrac

2 Apr 14, 2022
MOpt-AFL provided by the paper "MOPT: Optimized Mutation Scheduling for Fuzzers"

MOpt-AFL 1. Description MOpt-AFL is a AFL-based fuzzer that utilizes a customized Particle Swarm Optimization (PSO) algorithm to find the optimal sele

172 Dec 18, 2022
Implementation of Neural Distance Embeddings for Biological Sequences (NeuroSEED) in PyTorch

Neural Distance Embeddings for Biological Sequences Official implementation of Neural Distance Embeddings for Biological Sequences (NeuroSEED) in PyTo

Gabriele Corso 56 Dec 23, 2022
Example how to deploy deep learning model with aiohttp.

aiohttp-demos Demos for aiohttp project. Contents Imagetagger Deep Learning Image Classifier URL shortener Toxic Comments Classifier Moderator Slack B

aio-libs 661 Jan 04, 2023
SparseInst: Sparse Instance Activation for Real-Time Instance Segmentation, CVPR 2022

SparseInst 🚀 A simple framework for real-time instance segmentation, CVPR 2022 by Tianheng Cheng, Xinggang Wang†, Shaoyu Chen, Wenqiang Zhang, Qian Z

Hust Visual Learning Team 458 Jan 05, 2023
Home for cuQuantum Python & NVIDIA cuQuantum SDK C++ samples

Welcome to the cuQuantum repository! This public repository contains two sets of files related to the NVIDIA cuQuantum SDK: samples: All C/C++ sample

NVIDIA Corporation 147 Dec 27, 2022
Benchmarks for semi-supervised domain generalization.

Semi-Supervised Domain Generalization This code is the official implementation of the following paper: Semi-Supervised Domain Generalization with Stoc

Kaiyang 49 Dec 10, 2022
[NeurIPS'21] Shape As Points: A Differentiable Poisson Solver

Shape As Points (SAP) Paper | Project Page | Short Video (6 min) | Long Video (12 min) This repository contains the implementation of the paper: Shape

394 Dec 30, 2022
BasicVSR++: Improving Video Super-Resolution with Enhanced Propagation and Alignment

BasicVSR++: Improving Video Super-Resolution with Enhanced Propagation and Alignment

Holy Wu 35 Jan 01, 2023
Combining Automatic Labelers and Expert Annotations for Accurate Radiology Report Labeling Using BERT

CheXbert: Combining Automatic Labelers and Expert Annotations for Accurate Radiology Report Labeling Using BERT CheXbert is an accurate, automated dee

Stanford Machine Learning Group 51 Dec 08, 2022
Codebase for testing whether hidden states of neural networks encode discrete structures.

structural-probes Codebase for testing whether hidden states of neural networks encode discrete structures. Based on the paper A Structural Probe for

John Hewitt 349 Dec 17, 2022
PyTorch implementation of convolutional neural networks-based text-to-speech synthesis models

Deepvoice3_pytorch PyTorch implementation of convolutional networks-based text-to-speech synthesis models: arXiv:1710.07654: Deep Voice 3: Scaling Tex

Ryuichi Yamamoto 1.8k Jan 08, 2023
Unleashing Transformers: Parallel Token Prediction with Discrete Absorbing Diffusion for Fast High-Resolution Image Generation from Vector-Quantized Codes

Unleashing Transformers: Parallel Token Prediction with Discrete Absorbing Diffusion for Fast High-Resolution Image Generation from Vector-Quantized C

Sam Bond-Taylor 139 Jan 04, 2023
(ImageNet pretrained models) The official pytorch implemention of the TPAMI paper "Res2Net: A New Multi-scale Backbone Architecture"

Res2Net The official pytorch implemention of the paper "Res2Net: A New Multi-scale Backbone Architecture" Our paper is accepted by IEEE Transactions o

Res2Net Applications 928 Dec 29, 2022
STRIVE: Scene Text Replacement In Videos

STRIVE: Scene Text Replacement In Videos Dataset Types: RoboText SynthText RealWorld videos RoboText : Videos of texts collected using navigation robo

15 Jul 11, 2022
Accelerated SMPL operation, commonly used in generate 3D human mesh, STAR included.

SMPL2 An enchanced and accelerated SMPL operation which commonly used in 3D human mesh generation. It takes a poses, shapes, cam_trans as inputs, outp

JinTian 20 Oct 17, 2022
An index of algorithms for learning causality with data

awesome-causality-algorithms An index of algorithms for learning causality with data. Please cite our survey paper if this index is helpful. @article{

Ruocheng Guo 2.3k Jan 08, 2023