Home for "How To Scale Your Model", a short blog-style textbook about scaling LLMs on TPUs

AndroidWorld is an environment and benchmark for autonomous agents

Minimal reproduction of DeepSeek R1-Zero

A zero-dependency ML framework in C with a modern Python API for full control over execution and memory.

Evaluating Safety of Autonomous Agents in Mobile Device Control (AAAI 2026 AI Alignment Track)

ScreenAgent: A Computer Control Agent Driven by Visual Language Large Model (IJCAI-24)

[ICLR 2025] A trinity of environments, tools, and benchmarks for general virtual agents

[NeurIPS 2024] OSWorld: Benchmarking Multimodal Agents for Open-Ended Tasks in Real Computer Environments

Benchmarking Mobile Device Control Agents across Diverse Configurations (ICLR 2024 workshop GenAI4DM spotlight presentation; CoLLAs 2025)

Code for Paper: Autonomous Evaluation and Refinement of Digital Agents [COLM 2024]

Clean, modern, Python 3.6+ code generator & library for Protobuf 3 and async gRPC

Automatically manage all your ANC, MIR, SPEC token rewards, Mirror Delta Neutral Short Positions, Liquidity Pools, UST claims after lockup, Anchor Borrow and Earn for Terra.

Automating LTV Percentage

Python scripts for ETL (extract, transform and load) jobs for Ethereum blocks, transactions, ERC20 / ERC721 tokens, transfers, receipts, logs, contracts, internal transactions. Data is available in…

Reading list for research topics in multimodal machine learning

Official repository for the "Big Transfer (BiT): General Visual Representation Learning" paper.

bsuite is a collection of carefully-designed experiments that investigate core capabilities of a reinforcement learning (RL) agent

SEED RL: Scalable and Efficient Deep-RL with Accelerated Central Inference. Implements IMPALA and R2D2 algorithms in TF2 with SEED's architecture.

TensorFlow code and pre-trained models for BERT

A tiny scalar-valued autograd engine and a neural net library on top of it with PyTorch-like API

A face recognition for simpsons character build with flax

Trax — Deep Learning with Clear Code and Speed

Flax is a neural network library for JAX that is designed for flexibility.

Code for the paper "Generative Adversarial Imitation Learning"

Composable transformations of Python+NumPy programs: differentiate, vectorize, JIT to GPU/TPU, and more