Ninfer: High-performance single-GPU inference
NInfer offers a high-performance C++/CUDA inference engine for RTX 5090.
NInfer is a C++/CUDA inference engine designed to operate on a single NVIDIA GeForce RTX 5090. It supports specific Qwen3.6 checkpoints, processing text, image, and video prompts. Performance metrics were recorded with 7,680 and 260,096 token inputs, providing engineers with an optimized inference process.
This synthesis was produced from its source by AI; there is no human editor or manual review step. How we work