NeuroAdapter

Text-to-anatomical-map adapter checkpoint for NeuroVLM.

Files

  • model.safetensors: tensor weights for neurovlm.adapter.InterleavedDecoderAdapter.
  • config.json: adapter architecture and dependency metadata.

Load with:

from neurovlm.retrieval_resources import _load_neuro_adapter

adapter = _load_neuro_adapter(device="cpu")

Revision 2026-10

Weights of the revised NeuroVLM paper. What changed:

  • Deterministic training. Every stage runs with seeded data loaders and deterministic kernels. Two runs from scratch gave byte-identical weights on the same GPU model, driver and library versions.
  • MLP projector. The brain-to-text projectors (stages 13 and 14) are residual MLPs (28.3 M parameters) instead of Q-Formers. A package names its projector in config.json["projector"].
  • Stage-14 recipe. Each training text is projected onto the canonical bank of its own level, every training map is re-binarized per batch, and title tokens weigh 5 in the loss. Five seeds were trained and the one with the lowest validation loss (seed 3) is packaged.
  • Calibration. Text-to-brain maps are multiplied by one global factor per model, fitted on the validation split and clipped at 1 (map_calibration.json, head 19.16 and adapter 6.13).
  • Title decoding. After a [NETWORK] or [REGION] tag the decoder writes one of the canonical names in title_vocabulary.json (48 network and 83 region titles), so it never invents a title.

model.safetensors is the retrained anatomically centered projection head (stage 12, anchored recipe), and config.json is unchanged. Its calibration factor is adapter in map_calibration.json.

file size sha256
model.safetensors 135.0 MB 60d8ae78086462f1c261419dea0ce4409583c8e9cc8766195635abc80fd78372
map_calibration.json 0.9 kB 6f9daa11efb35446e32ad5604c36bdeae19ea3bc305cd5c3f289aea28154be83

Provenance: scripts/outputs/runs/final of NeuroVLM (scripts/pipeline/train_pipeline.py, finished 2026-10-10T19:56:25), training code at git 51bedf4 with 131 modified files, inputs pinned by scripts/config/hf_manifest.json (sha256 49f3e3a9825b). scripts/README.md records the later runs from scratch that gave the same bytes.

Earlier weights: revision 1db5f18c512dad7af14c35a548593557a9399f31.

Downloads last month
11
Safetensors
Model size
33.7M params
Tensor type
F32
ยท
Inference Providers NEW
This model isn't deployed by any Inference Provider. ๐Ÿ™‹ Ask for provider support