rtx

Definition

NVIDIA's consumer GPU line. When a paper or repository names an RTX card as its hardware, the claim being made is accessibility: the method runs on equipment an individual can actually buy, rather than on datacenter accelerators. The limiting factor is almost always VRAM rather than raw throughput, which is why work targeting these cards leans on quantization, offloading, and parameter-efficient fine-tuning.

Explainers using this term

• Paper walkthrough: Puro-2B — pretraining a 2B model from scratch for $6.9K on consumer GPUsPuro-2B: Poor Lab's Qwen2-1.5B Trained on RTX 5090 within $5090
• Paper Walkthrough: TurboVLA — Kick the LLM Out of the Loop and Run a Robot Policy at 32 Hz on an RTX 4090 with Under 1 GB of VRAMTurboVLA: Real-Time Vision-Language-Action Model at 32 Hz on an RTX 4090 with <1 GB VRAM