Official Codebase for "Neural Thickets: Diverse Task Experts Are Dense Around Pretrained Weights"

ARIS ⚔️ (Auto-Research-In-Sleep) — Lightweight Markdown-only skills for autonomous ML research: cross-model review loops, idea discovery, and experiment automation. No framework, no lock-in — works…

[ICML 2025] The Diffusion Duality

Scaling Beyond Masked Diffusion Language Models

The best ChatGPT that $100 can buy.

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

SDAR (Synergy of Diffusion and AutoRegression), a large diffusion language model（1.7B, 4B, 8B, 30B）

[ICLR 2026] Official code for TraceRL: Revolutionizing post-training for Diffusion LLMs, powering the SOTA TraDo series.

dLLM: Simple Diffusion Language Modeling

GSM8K-V: Can Vision Language Models Solve Grade School Math Word Problems in Visual Contexts

A nix configuration for chameleon

诺亚盘古大模型研发背后的真正的心酸与黑暗的故事。

Official Implementation for the paper "d1: Scaling Reasoning in Diffusion Large Language Models via Reinforcement Learning"

[ICLR 2025] Diffusion Feedback Helps CLIP See Better

Code for the Interspeech 2021 paper "AST: Audio Spectrogram Transformer".

Concept Bottleneck Models, ICML 2020

[ICLR'25 Oral] Representation Alignment for Generation: Training Diffusion Transformers Is Easier Than You Think

Official PyTorch Implementation of "Scalable Diffusion Models with Transformers"

Multi-lingual large voice generation model, providing inference, training and deployment full-stack ability.

Foundational model for human-like, expressive TTS

从0到1构建一个MiniLLM (pretrain+sft+dpo实践中)

[ICLR 2024] Fine-tuning LLaMA to follow Instructions within 1 Hour and 1.2M Parameters

deep learning for image processing including classification and object-detection etc.

Implementation of paper - YOLOv9: Learning What You Want to Learn Using Programmable Gradient Information