Repository Index

[SIGGRAPH Asia 2026] UniMate: One Unified Model to Animate Diverse Skeletons

Claude Code skill：把内容套进内置手绘画风配方,产出可直接复制的生图提示词。内置儿童涂色/极简线条/蜡笔童涂/吉卜力/小豆人涂鸦 5 种已验证画风。

Pytorch framework for doing deep learning on point clouds.

• pytorch
• deep-learning
• point-cloud
• pointnet

Serve large Qwen models fast on the GPUs you actually own. Qwen3.8-27B on a single 24 GB card with vLLM: 127 tok/s single-user (381 when the answer quotes the prompt), ~1,035 tok/s at 64 concurrent, 150k-262k context. vLLM patches, requant pipeline, benchmarks.

• kv-cache
• llm-inference
• local-llm
• quantization

Multimodal Unsupervised Image-to-Image Translation

• gan
• deep-learning
• pix2pix
• image-translation

Self-hosted LLM router with a managed safety net. OpenAI-compatible. BYOK. Single-workspace. Streaming. For more advanced routing choose hosted OrcaRouter

• ai-gateway
• ai-gateways
• anthropic-api
• byok

Pretrained DeepLabv3 and DeepLabv3+ for Pascal VOC & Cityscapes

• deeplabv3plus
• deeplabv3
• pytorch
• pascal-voc

Cross-platform CLI and Python drivers for AIO liquid coolers and other devices

• corsair
• evga
• nzxt
• liquid-cooler

SkyReels V1: The first and most advanced open-source human-centric video foundation model

• i2v
• t2v
• video-diffusion-transformers

Just draw a bounding box and you can remove the object you want to remove.

• pytorch
• inpainting
• video-inpainting
• image-inpainting

日本电影元数据刮削器，配合kodi,emby,plex等本地媒体管理工具使用。可批量抓取，也可单个抓取。可抓取子目录下视频，多集视频（-cd1/-cd2）,带字幕作品（-c., -C.）。批量添加emby演员头像。

• jav
• jav-scraper
• emby
• kodi

The fastest way to run Qwen 3.8 Flash Next, Qwen 3.8 27B and Ternary Bonsai 2 27B on a Mac: 125 tok/s in OpenCode on an M5 Max, and a 27B model on 16 GB Macs. Native MTP speculative decoding on Apple Silicon, exact at any temperature. OpenAI and Anthropic compatible local server.

• apple-silicon
• local-ai
• mlx
• mtp

Semi Supervised Learning for Medical Image Segmentation, a collection of literature reviews and code implementations.

• semi-supervised-learning

Crankshaft: A turnkey GNU/Linux solution that transforms a Raspberry Pi to an Android Auto head unit.

• raspberry-pi
• crankshaft
• rpi
• rpi3

EVA Series: Visual Representation Fantasies from BAAI

• foundation-models
• representation-learning
• vision-transformer

ASC is a super FAST and lightweight Android decompiler designed for Agents/Mobile Researchers.

• android
• decompilation
• decompile
• decompiler

[ICCV2025] LHM: Large Animatable Human Reconstruction Model from a Single Image in Seconds

• aicg
• aigc
• digitalhuman

Python with braces. Because python is awesome, but whitespace is awful.

• python
• whitespace
• braces
• indentation

Lightweight Python library for adding real-time multi-object tracking to any detector.

• tracking
• tracking-algorithm
• kalman-filter
• object-tracking

Utilize misconfigured DNS and old database records to find hidden IP's behind the CloudFlare network

• cloudflare
• tor
• ip
• database

Deep neural network to extract intelligent information from invoice documents.

• invoice
• invoice-management
• invoices
• invoice-insight

Auto-bypass for blocked websites in Claude Code — Phase 0→3 adaptive scheduler, no API keys

streamline the fine-tuning process for multimodal models: PaliGemma 2, Florence-2, and Qwen2.5-VL

• captioning
• fine-tuning
• florence-2
• multimodal

High-Quality Human Motion Video Generation with Confidence-aware Pose Guidance

• diffusion-models
• video-generation