Implement a reasoning LLM in PyTorch from scratch, step by step

Awesome Knowledge Distillation

A unified inference and post-training framework for accelerated video generation.

TurboDiffusion: 100–200× Acceleration for Video Diffusion Models

Awesome Knowledge-Distillation. 分类整理的知识蒸馏paper(2014-2021)。

推荐/广告/搜索领域工业界经典以及最前沿论文集合。A collection of industry classics and cutting-edge papers in the field of recommendation/advertising/search.

The official repo for [NeurIPS'22] "ViTPose: Simple Vision Transformer Baselines for Human Pose Estimation" and [TPAMI'23] "ViTPose++: Vision Transformer for Generic Body Pose Estimation"

PyTorch implementation of various methods for continual learning (XdG, EWC, SI, LwF, FROMP, DGR, BI-R, ER, A-GEM, iCaRL, Generative Classifier) in three different scenarios.

Pytorch implementation of various Knowledge Distillation (KD) methods.

A PyTorch-based knowledge distillation toolkit for natural language processing

PaddleSlim is an open-source library for deep model compression and architecture search.

All-in-one training for vision models (YOLO, ViTs, RT-DETR, DINOv3): pretraining, fine-tuning, distillation.

Reinforcement Learning via Self-Distillation (SDPO)

Generate High-Quality Synthetics, Train, Measure, and Evaluate in a Single Pipeline

mobilev2-yolov5s剪枝、蒸馏，支持ncnn，tensorRT部署。ultra-light but better performence！

高质量中文预训练模型集合：最先进大模型、最快小模型、相似度专门模型

♾️ 开源数字永生框架 — 从聊天记录蒸馏任何人的七维数字分身。支持微信/飞书/iMessage/Telegram等12+平台，7种角色模板，对齐 OpenClaw Soul Spec 标准。一行指令让你的AI学会蒸馏。

NVIDIA FastGen: Fast Generation from Diffusion Models

Kandinsky 5.0: A family of diffusion models for Video & Image generation

[ICML 2026] Official codebase for "Causal Forcing: Autoregressive Diffusion Distillation Done Right for High-Quality Real-Time Interactive Video Generation" & Causal Forcing++