AI & ML interests

Computer-use agents, typed decisions, on-device (MLX)

Recent Activity

gxcsoccer  updated a model about 1 month ago
deskmind/eyes-4b
gxcsoccer  updated a model about 1 month ago
deskmind/brain-4b
gxcsoccer  updated a model about 1 month ago
deskmind/brain-0.8b
View all activity

gxcsoccer 
posted an update about 1 month ago
view post
Post
108
I wrote up how DeskMind's local decision models work: a 0.8B and a 4B (Qwen3.5 fine-tunes, 8-bit MLX) that answer each desktop step as a multiple-choice question with a probability for every option. The 0.8B answers first, unsure steps go to the 4B, and when a task is ambiguous it asks the user instead of guessing.

The article also has three lessons from about twenty training rounds, including how smoothing the labels to 0.95 squeezed the 0.8B's confidence right onto the routing threshold.

Article: https://huggingface.co/blog/gxcsoccer/deskmind-local-decision-models
Models:
deskmind
  • 1 reply
·