Cuda agents

FLAME GPU 2 is a GPU accelerated agent based modelling framework for CUDA C++ and Python

RLVR training for LLM in CUDA/C++

From-scratch C++/CUDA inference engine for the NVIDIA RTX 5090 (sm_120a) — the best single-GPU backend for agentic AI: tool calling, long-context loops, reasoning and concurrent sub-agents on top of the fastest single-stream decode on the 5090 (beats llama.cpp, at-or-ahead of vLLM on NVFP4). 100% written by Claude Code.

🎓The path to LLM mastery is paved with broken embeddings and resurrected gradients. (CHINESE NOTES)

Fastest MoE/LLM inference runtime for consumer and edge Blackwell GPUs. SN74 on Gittensor.