NeuroFixer

Install

pip install neurofixer

Core modules

• EncodingGate — feature calibration with gated residual flow.
• EncodingModule — latent-grid alignment and dilation modulation.
• FusionBridge — branch-normalized multi-scale feature fusion.
• NeuromodulationController — resource-aware fusion control.

Quick usage

python -m neurofixer.demo.run_demo --backbone cnn
python -m neurofixer.demo.run_demo --backbone vit

from neurofixer.models.neurofuser_basic import (
    build_neurofuser_basic_cnn,
    build_neurofuser_basic_vit,
)

Community direction

NeuroFixer is designed to grow as a compact research library for attention heads, fusion modules, token-mixing blocks, gating functions, and model-family demos. Researchers are encouraged to share lightweight PyTorch attention modules with clear tensor-shape documentation, tests, and citation.