Qwen3.8-Flash-Next (512-expert MoE, 256k context) on a single Intel Gaudi2: native SynapseAI + custom TPC kernels, DFlash2 speculative decoding, OpenAI/Anthropic-compatible API, Docker. 单卡 Gaudi2 原生推理，decode ~155 tok/s，prefill ~3.5k tok/s。
GitHub repository with 10 stars and 3 forks.
Language: C++
Topics: docker, gaudi2, habana, intel-gaudi, llm-inference, moe, qwen, speculative-decoding, tpc, synapseai
Trending score 0.37, freshness score 0.08, stars gained +1, forks gained +0.
2026-11-02: 10 stars and 3 forks.
Qwen3.8-Flash-Next on any consumer hardware: one-click install for Windows / Linux. Strata inference engine, OpenAI/Anthropic API on localhost, optional image input.
GitHub repository with 38,584 stars and 4,196 forks.
Trending score: 5.97; stars gained: +1,346; forks gained: +175.
Language: C++
Tool for automatic PS5 executables porting to Linux and Windows
GitHub repository with 50,745 stars and 4,838 forks.
Trending score: 5.32; stars gained: +580; forks gained: +82.
Language: C++
Topics: anyps5, dynamic-library, game-porting, ps5, ps5-tools, spir-v
LLM inference in C/C++
GitHub repository with 133,212 stars and 25,073 forks.
Trending score: 4.62; stars gained: +206; forks gained: +59.
Language: C++
Topics: ggml
EditHere — 截图、写下意见、直接调整布局，把修改意图一次交给 AI。
GitHub repository with 1,258 stars and 88 forks.
Trending score: 3.91; stars gained: +90; forks gained: +4.
Language: C++
An MCP-based chatbot | 一个基于MCP的聊天机器人
GitHub repository with 32,393 stars and 7,623 forks.
Trending score: 3.84; stars gained: +58; forks gained: +17.
Language: C++
Topics: chatbot, esp32, mcp
The Goose Programming Language
GitHub repository with 2,868 stars and 117 forks.
Trending score: 4.45; stars gained: +249; forks gained: +8.
Language: C++
Opinionated PHP Docker images
GitHub repository with 43 stars and 6 forks.
Trending score: 0.00.
Language: PHP
Topics: docker, hacktoberfest, parallel, php, php7, php8