samlobel

samlobel.github.io

Jekyll webage

RaCT_CF

Ranking-Critical Training for Collaborative Filtering

cleanrl_starter

CleanRL set up with logging etc

many_gamma

What do you do with value estimates at a variety of gammas?

CFN

Accompanying Code for "Flipping Coins to Estimate Pseudocounts for Exploration in Reinforcement Learning", ICML 2023

grl

InfinitePhoto

A picture within a picture within picture

spelling_bee

A little script to get the highest-scoring set of letters in NYTimes spelling bee

launchpad

onager

Lightweight python library for launching experiments and tuning hyperparameters, either locally or on a cluster

q_functionals

Accompanying code for "Q-Functionals for Value-Based Continuous Control"

dopamine

Dopamine is a research framework for fast prototyping of reinforcement learning algorithms.

optimistic-init

Accompanying code for "Optimistic Initialization for Exploration in Continuous Control"

adxgame

Implementation of the AdxGame.

pfrl

PFRL: a PyTorch-based deep reinforcement learning library

mrl

gym

A toolkit for developing and comparing reinforcement learning algorithms.

rl_scaffolding

A framework structured the way I like, that I can branch off of for new experiments

GAILA

skill-chaining

Implementation of "Skill Discovery in Continuous Reinforcement Learning Domains using Skill Chaining"

RBFDQN

grad-cnns

Code and manuscript for "Efficient Per-Example Gradient Computations in Convolutional Neural Networks"

simple_rl_mbrl

Simple-RL forked MBRL

torchbeast

A PyTorch Platform for Distributed RL

Torchelie

Torchélie is a personal set of utility functions and layers for PyTorch.

simple_rl

A simple framework for experimenting with Reinforcement Learning in Python.

rpc4django

RPC4Django is an XMLRPC and JSONRPC server for Django powered projects. Simply plug it into any existing Django project and you can make your methods available via XMLRPC and JSONRPC. This project is actively seeking maintainers.

spider-sim

Simulating a Spider catching its prey

SQL-Rank

A Novel Listwise Collaborative Ranking Algorithm published at ICML'18

pytorch-utils

Code snippets in PyTorch I've found useful