365 Days Computer Vision Learning Linkedin Post

Code for the model presented in the paper: "code2seq: Generating Sequences from Structured Representations of Code"

This repository contains a Pytorch implementation of the paper "The Lottery Ticket Hypothesis: Finding Sparse, Trainable Neural Networks" by Jonathan Frankle and Michael Carbin that can be easily adapted to any model/dataset.

Lanczos Network, Graph Neural Networks, Deep Graph Convolutional Networks, Deep Learning on Graph Structured Data, QM8 Quantum Chemistry Benchmark, ICLR 2019

PyTorch code for ICLR 2019 paper: Self-Monitoring Navigation Agent via Auxiliary Progress Estimation

[ICLR'19] Meta-learning with differentiable closed-form solvers

Official PyTorch implementation of Harmonizing Maximum Likelihood with GANs for Multimodal Conditional Generation (ICLR 2019)

A simplified PyTorch implementation of GANsynth

[ICLR'19] Complement Objective Training

The Reinforcement-Learning-Related Papers of ICLR 2019

Variance Networks: When Expectation Does Not Meet Your Expectations, ICLR 2019

Code for the paper 'Neural Persistence: A Complexity Measure for Deep Neural Networks Using Algebraic Topology'

✂️ Repository for our ICLR 2019 paper: Discovery of Natural Language Concepts in Individual Units of CNNs

PyTorch implementation of "Variational Autoencoders with Jointly Optimized Latent Dependency Structure" [ICLR 2019]

Single shot neural network pruning before training the model, based on connection sensitivity

ICLR 2020 and 2019 reviews

We propose a Seed-Augment-Train/Transfer (SAT) framework that contains a synthetic seed image dataset generation procedure for languages with different numeral systems using freely available open font file datasets

Implementation of https://arxiv.org/pdf/1805.12352.pdf (ICLR 2019)