SNN
Applied AI Engineer Experimental Access on request
Updated: 2026-02-18
Research bench for toy SNNs, Lexi memory-core scaffolds, and SpikingBrain-7B integration experiments.
Impact: Keeps long-horizon neuromorphic memory ideas in a bounded lab lane separate from production assistant claims.
What I built
- Runs toy snnTorch demos for spike dynamics, surrogate-gradient training, and temporal memory exploration.
- Includes a Lexi memory-core scaffold for event and embedding streams with affective state experiments.
- Documents optional SpikingBrain-7B weight placement and a minimal CLI harness with clear fallback guidance.
Proof: Run `./scripts/run_toy_snn_demo.sh` and `./scripts/run_lexi_memory_demo.sh` from the README.
PythonPyTorchsnnTorchTransformersModelScope