Havuz

:sunrise: A comprehensive, cutting-edge open-source framework for next-generation Transformers/LLM, with compute-matched benchmarking and a

A high-throughput training framework for Speculative decoding models.

Halo is an open-source framework built by White Circle for training large language and multimodal models

🚀🚀 A multimodal System One decision model that gives calibrated answers to typed questions about screens, photos, video and text in one forw

Refusal removal as a loadable KV-cache graft — zero weight modification, no refusal-direction projection, fully reversible.

Single-GPU LLM fine-tuning preflight: memory estimates, runnable LoRA/QLoRA recipes, and local measurements.

Free Course : Data, Machine Learning, Deep Learning and AI Engineering, the long way round: the math, the code, the systems and the security

Bilingual (中文+EN) ML / LLM / diffusion / agent interview cheat sheets for AI 秋招 — generated by ARIS /interview-cheatsheet, rendered by /rend

From teacher to tiles — a from-scratch LLM distillation & serving engine: custom Triton/CUDA kernels, FSDP distillation, paged-KV continuous

Local decision engine with typed, calibrated outputs (Noul/Choice/Score): cheap→strong cascade, ECE/Brier calibration, SystemOne protocol in

Transformers for decision models: fine-tune, load and serve calibrated decsion-model on any Hugging Face model.

Jacobian-Brainwash : A manual alignment tool for large language models built on Anthropic's Jacobian Lens. Results are exportable.

🔥🔥 Accelerating decision inference with exact state forking

To know what models don't say out loud.

从 MiniMind 源码读起,再延伸到现代大模型技术体系的中文学习笔记。主线逐行精读预训练 / SFT / DPO / PPO / GRPO 与训练机制;附录 17 篇进阶卷覆盖量化、投机解码、RLHF 全景、模型代际史等 MiniMind 没涉及、但进阶绕不开的主题。

Scaler 2 Music Theory Tool 2026: Direct DAW Integration for Windows 10 and 11

PyTorch 기반 한국어 LLM 구현 및 학습 프로젝트

This repo is a alternative solution for standard self-attention. It uses tree routing with learnable functions for cheaper self-attention wh

Best AI Workload Orchestrator for Kubernetes Python 2026

Automated alignment adjustment for LLMs — direct steering, LoRA, and MoE expert-granular abliteration, optimized via multi-objective Optuna

Streamline AI Workloads on Kubernetes: Universal Python SDK Guide 2026

Windows utility focused on cuda setup, nvidia local AI, and cuda environment for local AI workflows.

Lev: Thinking Fast and Deep with Looped Decision Models. One checkpoint, multiple depths: System 1.5 decision models that think in loops, no

A curated collection of AI learning materials

Self-Tuning Multi-Agent AI Trading System 2026: 8-Source Signal Fusion & Kronos Model

Open, calibrated System 1 decision models: typed choice / score / yes-no answers with a probability for every option. 400M to 80B; CPU, NVID

A small chat model trained from scratch: tokenizer → pretraining → SFT → RL → multimodal. 0.78B params, CORE 0.248, tool calling, vLLM. Weig

Train JEV-style System 1 decision models on Qwen3.5 and Qwen3.8-27B: 24-slot decision head, KL distillation, calibration and serving. 在 Qwen

The htop for LLM inference see exactly where every GB of VRAM goes and get measured quantization savings.

Building a small LLM in PyTorch from scratch - a step-by-step journey inspired by Stanford CS336.

Learn Agentic AI, Deep Learning, Agentic Engineering, RAG and Multi-Agent Workflows.

GleamLM大语言模型全栈训练平台——核心代码用 PyTorch 从零实现，最低仅需 12GB 消费级显卡即可完整训练，另有一套工业代码与手写实现消费同一份数据逐层对照。模型训练支持CLI和可视化训练。

Composable LLM decoding on the probability simplex: library and paper code.

🌹 Rose: Range-Of-Slice Equilibration PyTorch optimizer. Stateless optimization through range-normalized gradient updates.

Build world models from scratch — from perception and latent dynamics to planning, JEPA, VLA, and 3D/4D intelligence.

First open-source implementation of Google TurboQuant (ICLR 2026) -- near-optimal KV cache compression for LLM inference. 5x compression wit

Pyramis-L3D: hierarchical sparse attention research prototype (code only; weights on Hugging Face)

Production-ready, bleeding-edge, upstream-first vLLM Docker image for the NVIDIA DGX Spark (GB10 / sm_121a), verified against real models on

AnyFlow - flow-matching video generation: training, inference and VBench evaluation for t2v and i2v models.

Self-tuning multi-agent AI trading system. 8-source signal fusion, Bull/Bear/Judge debate on Claude Opus 4.7, Kelly + ATR position sizing. P

Evaan — a fully local, offline-capable AI companion chatbot built with Python & Hugging Face Transformers, running on Qwen2.5-0.5B-Instruct

Inference engine for large language models on consumer GPUs, with deep optimization for older hardware

Open-source Environment toolkit of claw-like agents, support task/harness generation and evaluation

🚀 Engram-PEFT: An unofficial implementation of DeepSeek Engram. Inject high-capacity conditional memory into LLMs via sparse retrieval PEFT

Implementation of LatentMoE,Toward Optimal Accuracy per FLOP and Parameter in Mixture of Experts (Elango et al., NVIDIA 2026) — in Pytorch.

🔥 AI agent that ports CUDA code to AMD ROCm using local LLMs — FastAPI + React + Ollama | Live: rocmporter-agent.vercel.app

a LLM inference engine to run on consumer hardware

One GPU. Full LLM workflow. Real benchmarks. No cloud required.

First open-source TurboQuant KV cache compression for LLM inference. Drop-in for HuggingFace. pip install turboquant.

A tiny language model that talks like a house cat named Miso.