JunkyByte
Follow
Lists (1)
Stars
[CVPR 2026] Official code and models for Video Encoder-only Mask Transformer (VidEoMT).
The repository provides code for running inference with the SAM 3D Body Model (3DB), links for downloading the trained model checkpoints and datasets, and example notebooks that show how to use the…
MetaSlot: Break Through the Fixed Number of Slots in Object-Centric Learning
CLI proxy that reduces LLM token consumption by 60-90% on common dev commands. Single Rust binary, zero dependencies
[CVPR 2026 Oral] "INSID3: Training-Free In-Context Segmentation with DINOv3"
An Obsidian plugin to open files in Cursor IDE or other VSCode-based editors (VSCode, Kiro, etc.) with automatic cursor position jumping. 一个可以在 Cursor IDE 或其他类 VSCode 编辑器中打开文件并自动跳转光标位置的 Obsidian 插件。
DiffusionBlocks: Block-wise Neural Network Training via Diffusion Interpretation
Reinforcement Learning via Self-Distillation (SDPO)
The repository provides code for running inference and finetuning with the Meta Segment Anything Model 3 (SAM 3), links for downloading the trained model checkpoints, and example notebooks that sho…
AHN: Artificial Hippocampus Networks for Efficient Long-Context Modeling
A concise but complete full-attention transformer with a set of promising experimental features from various papers
Wan: Open and Advanced Large-Scale Video Generative Models
Portable file server with accelerated resumable uploads, dedup, WebDAV, SFTP, FTP, TFTP, zeroconf, media indexer, thumbnails++ all in one file
Hierarchical Reasoning Model Official Release
Mirage Persistent Kernel: Compiling LLMs into a MegaKernel
Enjoy puzzle-solving directly in your browser.
Official codebase for "Self Forcing: Bridging Training and Inference in Autoregressive Video Diffusion" (NeurIPS 2025 Spotlight)
ConceptSearch: Towards efficient program search using LLMs for Abstraction and Reasoning Corpus (ARC)
PyTorch implementation of the InfoNCE loss for self-supervised learning.
Solving the Abstraction & Reasoning Corpus with DreamCoder