jochi2018/Soup: Fine-tune and post-train LLMs with one command—no SSH, no config hell.
Fine-tune and post-train LLMs with one command—no SSH, no config hell.
Repositories
bawadou/ai-data-extractor: Free open-source extractor for AI coding assistant chat histories. Supports Claude Code, Cursor, Windsurf, Aider, Cline/Roo Code, and more.
Free open-source extractor for AI coding assistant chat histories. Supports Claude Code, Cursor, Windsurf, Aider, Cline/Roo Code, and more.
shyringo/deepseek-v4-flash-0731-in-c: Run the native 284B-A13B DeepSeek-V4-Flash-0731 LLM locally on a single laptop CPU: pure C, 8 GB RAM minimum, no GPU, best TPOT 0.892 s/token. ｜ 在笔记本单颗 CPU 上本地运行原生 284B-A13B DeepSeek-V4-Flash-0731 大模型：纯 C，最低 8 GB 内存，无需 GPU，最优 TPOT 0.892 秒/token。
Run the native 284B-A13B DeepSeek-V4-Flash-0731 LLM locally on a single laptop CPU: pure C, 8 GB RAM minimum, no GPU, best TPOT 0.892 s/token. ｜ 在笔记本单颗 CPU 上本地运行原生 284B-A13B DeepSe...
K-Dense-AI/drug-discovery-agent-skills: Agent Skills for small-molecule and protein therapeutics: cheminformatics, molecular ML, docking and dynamics, protein design platforms, and target-discovery knowledge graphs.
Agent Skills for small-molecule and protein therapeutics: cheminformatics, molecular ML, docking and dynamics, protein design platforms, and target-discovery knowledge graphs.
lucidrains/bdh-cq: Implementation of BDH-CQ: In-Context Learning with Recurrent Latent Reasoning, proposed by Pathway Research
Implementation of BDH-CQ: In-Context Learning with Recurrent Latent Reasoning, proposed by Pathway Research
alphaparkinc/genpark-multimodal-generative-art-prompt-synthesis-engine-skill: Multimodal generative art prompt synthesis & diffusion renderer (SeaArt style)
Multimodal generative art prompt synthesis & diffusion renderer (SeaArt style)
Zhou-Shilin/SynthID-Text-Bypass: Research demo: scrub generative text watermarks (e.g. SynthID-Text) via H-token sliding-window MLM token replacement. Gradio UI, multilingual (ModernBERT / mmBERT).
Research demo: scrub generative text watermarks (e.g. SynthID-Text) via H-token sliding-window MLM token replacement. Gradio UI, multilingual (ModernBERT / mmBERT).
Krishita17/PhantomTap: A Flipper Zero that stops guessing and starts reasoning: ML-guided RFID/NFC fuzzing & access-control auditing that turns raw card reads into a prioritized, explainable security assessment. Defensive-use only.
A Flipper Zero that stops guessing and starts reasoning: ML-guided RFID/NFC fuzzing & access-control auditing that turns raw card reads into a prioritized, explainable security ass...
zjp1997720/zhijian-ai-bluebook-workbuddy-harness: 智见 AI 蓝皮书：拆解 WorkBuddy 的提示词、记忆、插件、专家、Skill 与安全边界。
智见 AI 蓝皮书：拆解 WorkBuddy 的提示词、记忆、插件、专家、Skill 与安全边界。
Greninja9257/LabLLM: A native macOS lab for teaching tiny language models to think — build the architecture, train the weights, and watch a small LLM emerge from scratch, locally on Apple Silicon with custom data, tokenizers, checkpoints, and MLX acceleration.
A native macOS lab for teaching tiny language models to think — build the architecture, train the weights, and watch a small LLM emerge from scratch, locally on Apple Silicon with ...
dickpy/dsh-imagegen: DSH (DeepSeek Harness) Web GUI AI image generation plugin: text-to-image & image-to-image via OpenAI-compatible endpoints (gpt-image-2), with shared cross-device history.
DSH (DeepSeek Harness) Web GUI AI image generation plugin: text-to-image & image-to-image via OpenAI-compatible endpoints (gpt-image-2), with shared cross-device history.
dinesshh2930/AI_Data_Analyst: AI-powered end-to-end data analytics application built with Python, Pandas, Streamlit and Generative AI.
AI-powered end-to-end data analytics application built with Python, Pandas, Streamlit and Generative AI.
umran666/caliper: High-precision, zero-dependency Byte-Fallback Unigram tokenizer with dual-offset tracking, arithmetic isolation, and multilingual Unicode protection.
High-precision, zero-dependency Byte-Fallback Unigram tokenizer with dual-offset tracking, arithmetic isolation, and multilingual Unicode protection.
DjDevilCloud/EmitterRefusion: Novel Text To Image Architecture
Novel Text To Image Architecture
jeremy-prt/bloub: SVG recreation of the x.ai bot avatar. One shape morphing through 14 states, measured off the reference video frame by frame.
SVG recreation of the x.ai bot avatar. One shape morphing through 14 states, measured off the reference video frame by frame.
fudan-generative-vision/WAM-Diff2: Hierarchical AR-to-Diffusion Distillation for Efficient Autonomous Driving VLA
Hierarchical AR-to-Diffusion Distillation for Efficient Autonomous Driving VLA
cneuralnetwork/oplogs: Local-first, open-source experiment tracking for machine learning and agents
Local-first, open-source experiment tracking for machine learning and agents
xiaozhe7772222/dsh-draw-router: Universal image generation router for DeepSeek Harness (DSH): auto-discovers image models from any OpenAI-compatible endpoint. Supports SenseNova, StepFun, Agnes, Qwen, Flux, SD, Imagen and more. | DeepSeek Harness 统一绘图路由器
Universal image generation router for DeepSeek Harness (DSH): auto-discovers image models from any OpenAI-compatible endpoint. Supports SenseNova, StepFun, Agnes, Qwen, Flux, SD, I...
umanamir659/generative-ai-chatbot: python based AI using openAI API
python based AI using openAI API
mbsdeepak/text-diffusion-fashion-mnist: Text-conditioned diffusion model (tiny Stable Diffusion) built from scratch in PyTorch, trained on Fashion-MNIST — CLIP text conditioning, DDIM sampling, classifier-free guidance.
Text-conditioned diffusion model (tiny Stable Diffusion) built from scratch in PyTorch, trained on Fashion-MNIST — CLIP text conditioning, DDIM sampling, classifier-free guidance.
jaredrhod/barehands: Move things on your screen with your bare hands. A webcam-powered, hand-tracked interface for your AI. No headset. No controllers.
Move things on your screen with your bare hands. A webcam-powered, hand-tracked interface for your AI. No headset. No controllers.
andrew-yw/algothon-2026: Insights from Team abc123 on our Susquehanna x UNSW Algothon 2026 strategy, research process, ablations, and causal backtesting.
Insights from Team abc123 on our Susquehanna x UNSW Algothon 2026 strategy, research process, ablations, and causal backtesting.
fufankeji/deepseek-harness-studio: DeepSeek Harness 零代码桌面端｜一键启动，支持 Windows 与 macOS；内置插件发现、热点插件推送、一键安装与管理、AI 智能推荐和视觉增强。
DeepSeek Harness 零代码桌面端｜一键启动，支持 Windows 与 macOS；内置插件发现、热点插件推送、一键安装与管理、AI 智能推荐和视觉增强。
soumatheusgomes/vibe-coding-toolkit: A curated, battle-tested AI-coding toolkit: Claude Code plugins, subagent orchestration, quality gates, and ready-to-copy prompts, extracted from real production use.
A curated, battle-tested AI-coding toolkit: Claude Code plugins, subagent orchestration, quality gates, and ready-to-copy prompts, extracted from real production use.