LLaDA2.0 is the diffusion language model series developed by InclusionAI team, Ant Group.

A unified, torch-native training framework for MTP and block-parallel speculative decoding.

Strong Stochastic Flow Maps.

VOICE → WORDS

Code and paper reproductions for MLPs Are Hebbians

Code for A2D2: Fine-Tuning Any-Length Discrete Diffusion for Adaptive Decoding

The Precursor Genome is a dataset of 1,035 pairwise solid-state (ceramic) synthesis reactions generated autonomously by the A-Lab.

Codebase for the paper "Rethinking On-Policy Self-Distillation for Thinking Models"

Free Draft-and-Verification: Toward Lossless Parallel Decoding for Diffusion Large Language Models

Code accompanying the paper "Generalized Interpolating Discrete Diffusion"

A Python DSL to write Nvidia PTX for Hopper and Blackwell in JAX and PyTorch

ROCmFPX Family for AMD Hardware and Processors. More quants and special agent quants

The Lean Machine Learning Library

A library for soft differentiable relaxations of common PyTorch functions.

TiRex-2: A Zero-Shot Timeseries Forecasting Model

TabFM (Tabular Foundation Model) is a pretrained tabular foundation model developed by Google Research for tabular data regression and classification.

[ICLR2026] Official Code for Loopholing Discrete Diffusion: Deterministic Bypass of the Sampling Wall

A Python port of Pi’s minimalist coding agent.

Codebase for TIME benchmark

TS-ICL: a Time-Indexed Foundation Model for Time Series Forecasting & Imputation via In-Context Learning

MulTaBench: A Multimodal Tabular Benchmark

An open-weights time-series forecasting foundation model from The Forecasting Company.

An LLM post-training framework with vLLM for RL Scaling

Model export recipes, Python primitives, and Swift runtime utilities for on-device AI