zils-lite-k3b
A body for the Zils d3-lite soup pool (Bittensor subnet 127): a fine-tune of Qwen/Qwen3.5-0.8B in the chassis layout, BF16, language weights trained, vision tower from the parent.
A gentle fine-tune of the current soup (thermopylae-777/zils-lite-body-2 and dasLOL/zils-lite-skills-v5, both Apache-2.0) with soft targets, computed questions and regularisation to the starting model; top layers only.
Training data: project-generated computed questions and public train splits under their own licences. Check the original datasets' terms before reuse.
Inference Providers NEW
This model isn't deployed by any Inference Provider. ๐ Ask for provider support