Ariel Kwiatkowski's repositories
anterion
Open-source software engineer
AutoGPT
An experimental open-source attempt to make GPT-4 fully autonomous.
cogment-verse
Research platform for Human-in-the-loop learning (HILL) & Multi-Agent Reinforcement Learning (MARL)
Cradle
The Cradle framework is a first attempt at General Computer Control (GCC). Cradle supports agents to ace any computer task by enabling strong reasoning abilities, self-improvment, and skill curation, in a standardized general environment with minimal requirements.
Gymnasium
A standard API for reinforcement learning and a diverse set of reference environments (formerly Gym)
instructor
structured outputs for llms
keras
Deep Learning for humans
laser
The Truth Is In There: Improving Reasoning in Language Models with Layer-Selective Rank Reduction
LaVague
Copilot for web automation
llm.c
LLM training in simple, raw C/CUDA
lm-evaluation-harness
A framework for few-shot evaluation of language models.
OSWorld
OSWorld: Benchmarking Multimodal Agents for Open-Ended Tasks in Real Computer Environments
Shimmy
An API conversion tool for popular external reinforcement learning environments
SWE-agent
SWE-agent: Agent Computer Interfaces Enable Software Engineering Language Models
task-standard
METR Task Standard
TinyLlama
The TinyLlama project is an open endeavor to pretrain a 1.1B Llama model on 3 trillion tokens.
TPU-Alignment
Fully fine-tune large models like Mistral, Llama-2-13B, or Qwen-14B completely for free
vimGPT
Browse the web with GPT-4V and Vimium
wildcats-ai
This will one day be an actually working AI agent