Geometric Language Models β Research Lineages
Collection
Explore our native spectral, rotor, and oscillator models: interactive demos, clean releases, and explicitly labeled research lineages. β’ 15 items β’ Updated
WireNative 2M β multi-domain training (philosophy + code + science)
Part of the Harmonic GPT research into oscillator-based neural computation.
| Property | Value |
|---|---|
| Parameters | 2,376,024 |
| BPB | 0.0000 |
| Training step | 0 |
| n_harmonics | 32 |
| n_layers | 8 |
| n_groups | 7 |
| d_model | 448 |
| Vocab | 256 (raw bytes) |
from huggingface_hub import hf_hub_download
from safetensors.torch import load_file
weights = load_file(hf_hub_download("MonumentalSystems/wire-2m-multidomain", "model.safetensors"))
config = json.load(open(hf_hub_download("MonumentalSystems/wire-2m-multidomain", "config.json")))
All operations are native Clifford algebra / harmonic oscillator dynamics β no softmax attention, no MLP, no ReLU.