Downloads · 30 days
0
fly-brain-grok
monomyth/fly-brain-grok
fly-brain-grok is a robotics model from monomyth. Use it for the robotics task on the model card, and read the license before you ship it in a product. It is set up for numpy. The card lists the license as mit.
Distilled MaleCNS crop LIF gains (g) plus a 12-parameter DN→TCP unpacker (U) for the ReBot B601-DM kinematic arm in monomyth/rebot-motion-lab (fly-brain-grok branch).
Access
Public
Updated Sep 13, 2026
Repo size
19.1 MB
Likes
0
Public
Repo makeup
Click a slice to open those files.
.npz19.1 MB · 99%
At a glance
- Task
- Robotics
- Library
- numpy
- License
- mit
- Access
- Public
- Created
- Sep 12, 2026
- Updated
- Sep 13, 2026
- SHA
- c0f3798c
Model card
From the Hugging Face model README
fly-brain-grok
Distilled MaleCNS crop LIF gains (g) plus a 12-parameter DN→TCP unpacker (U) for the ReBot B601-DM kinematic arm in monomyth/rebot-motion-lab (fly-brain-grok branch).
This is not a Transformers checkpoint. Load with numpy.load.
Code and run instructions: github.com/monomyth/fly-brain-grok
Files
| File | What |
|---|---|
g-distill.npz | g (per-class synaptic gains) and u (U params including w_contact) |
g-distill.json | Same vectors + g_hash 7bca9074a252951b |
hop-probe.json | Crop identity: 59,740 neurons, 267 scored DNs |
malecns-v1.0-crop-v1-optic-rich-mancType-DNfl-DNxl.npz | Optic+DN CSR crop (optional; skip full graph prepare) |
eval-size20.json | Last 20 mm live episode summary |
Load
import numpy as np
blob = np.load("g-distill.npz")
g, u = blob["g"], blob["u"] # u[-1] is w_contact
Pass g-distill.npz to controller/scripts/run_dn_bus.py --gains ....
Provenance
MaleCNS / MANC types from the public Janelia male CNS connectome. Overlay / k-NN teachers are not stored here. da_learned is false. After a successful hold the current client Folds without opening, so the cube is not released.
- Task
- Robotics
- Library
- numpy
- License
- mit
- Created
- Sep 12, 2026
- Updated
- Sep 13, 2026