Unsloth says NVIDIA RTX Spark lets users run and train models above 120B parameters locally on a laptop. The post says training diffusion, decision and multimodal models runs about 2.7 times faster with CUDA on ARM, with up to 128GB of unified memory. The attached chart compares RTX Spark N1X with a premium SoC, showing 2.7x for fine-tuning with Unsloth.
The post links a hardware launch to a specific software speedup and memory limit, which helps readers judge what local training on this laptop class can handle.
You can now run and train 120B+ parameter models locally on a laptop! NVIDIA RTX Spark is here. 💚
With Unsloth and RTX Spark, train diffusion, decision, multimodal models ~2.7× faster with optimized kernels and binaries for CUDA on ARM. All local with up to 128GB unified memory
NVIDIA RTX Spark is here. Build AI locally, bring your creative ideas to life, and take your favorite games along. A new beginning for @Windows PCs starts today.
Source: Unsloth AI · x.comPublished · added here
