Global

pth） - 支持跨不同 GPU 数量恢复（自动调整 step） - 支持 wandb 训练记录连续性（自动恢复同一个 run）

Distilly is a local-first product for turning a person's source material, working habits, judgment, and voice into a versioned Person Profile for Agents. The profile can be recalled temporarily during a run or explicitly installed as a long-lived host Skill.

The website provides visual Skill Pack browsing, a beginner-friendly usage guide, Skill detail pages, and a contribution submission entry.

DeepSeek-R1-Distill models are fine-tuned based on open-source models, using samples generated by DeepSeek-R1. We slightly change their configs and tokenizers. Please use our setting to run these models.

A unified library of SOTA model optimization techniques like quantization, distillation, pruning, neural architecture search, speculative decoding, etc. It compresses deep learning models for downstream deployment frameworks like TensorRT-LLM, TensorRT, vLLM, etc.

We are releasing MiniCPM5-2B, the second model in the MiniCPM5 series, following MiniCPM5-1B. It is a dense 2B Transformer that scales up the same training recipe, built for on-device, local deployment, and resource-constrained scenarios, reaching 2B-class open-source SOTA.

本项目旨在分享大模型相关技术原理以及实战经验（大模型工程化、大模型应用落地）

The largest collection of PyTorch image encoders / backbones.

《李宏毅深度学习教程》（李宏毅老师推荐👍，苹果书🍎），PDF下载地址：

This is an official implementation for "Swin Transformer: Hierarchical Vision Transformer using Shifted Windows".

py is the client-facing contract (commands, messages, tracks), and the streaming client speaks exactly that — through reactor-sdk, the Reactor Python SDK, which carries the session, the commands and messages, and the WebRTC media tracks.

IMPORTANT] > Seeking Research and Dev. Collaborations. We are actively looking for research > collaborators to explore downstream applications built on DMI such as interpretability, speculative decoding, > hallucination analysis, distillation, activation steering, and beyond.