Decision model reportedly trained in about two minutes on a single DGX Spark
A post says the model used Unsloth for training, with Qwen3.5-2B as its base model.
TLDR
A post describes a video of a decision model trained in about two minutes on a single DGX Spark using Unsloth. It identifies Qwen3.5-2B as the base model and links to Unsloth’s training instructions.
Combined views
—
1 Source, first seen ago
