PyTorch implementation of MAR+DiffLoss https://arxiv.org/abs/2406.11838

一个基于nano banana pro🍌的原生AI PPT生成应用，迈向真正的＂Vibe PPT＂; 支持上传任意模板图片；上传任意素材&智能解析；一句话/大纲/页面描述自动生成PPT；口头修改指定区域、一键导出 - An AI-native PPT generator based on nano banana pro🍌

StarVector is a foundation model for SVG generation that transforms vectorization into a code generation task. Using a vision-language modeling architecture, StarVector processes both visual and te…

Accelerating Large-Scale Reasoning Model Inference with Sparse Self-Speculative Decoding

[ASPLOS'26] Taming the Long-Tail: Efficient Reasoning RL Training with Adaptive Drafter

Code for the paper “Four Over Six: More Accurate NVFP4 Quantization with Adaptive Block Scaling”

Real-Time VLAs via Future-state-aware Asynchronous Inference.

[CVPR 2025] Diffusion-4K: Ultra-High-Resolution Image Synthesis with Latent Diffusion Models

OmniVinci is an omni-modal LLM for joint understanding of vision, audio, and language.

SD.Next: All-in-one WebUI for AI generative image and video creation

StreamingVLM: Real-Time Understanding for Infinite Video Streams

StreamDiffusion, Live Stream APP

Portable ComfyUI for Windows, macOS and Linux 🔹 Pixaroma Community Edition 🔹

ComfyUI Plugin of Nunchaku

Build mindmaps with plain text

DC-Gen: Post-Training Diffusion Acceleration with Deeply Compressed Latent Space

DC-VideoGen: Efficient Video Generation with Deep Compression Video Autoencoder

LongLive: Real-time Interactive Long Video Generation

Multi-Platform Package Manager for Stable Diffusion

Pip / uv installable ComfyUI package

RAG on Everything with LEANN. Enjoy 97% storage savings while running a fast, accurate, and 100% private RAG application on your personal device.

Python package for parsing configurations from YAML with the command-line interface.

Run your GitHub Actions locally 🚀

ICLR Points: How Many ICLR Publications Is One Paper in Each Area?

Long-RL: Scaling RL to Long Sequences (NeurIPS 2025)

Locality-aware Parallel Decoding for Efficient Autoregressive Image Generation

[ICML 2025] SparseLoRA: Accelerating LLM Fine-Tuning with Contextual Sparsity