Downloads · 30 days
0
lila-e8-fandom-multiverse
h4/lila-e8-fandom-multiverse
lila-e8-fandom-multiverse is a machine learning model from h4. Use it for the machine learning task on the model card, and read the license before you ship it in a product. The card lists the license as agpl-3.0.
Lila-E8 trained on Fandom Multiverse (Monster + Time2026 + LMFDB + MathJax)
Access
Public
Updated Mar 9, 2026
Repo size
501 MB
Likes
1
Public
Repo makeup
Click a slice to open those files.
.pt501 MB · 100%
At a glance
- License
- agpl-3.0
- Access
- Public
- Created
- Mar 9, 2026
- Updated
- Mar 9, 2026
- SHA
- df623b4a
Model card
From the Hugging Face model README
lila-e8-fandom-multiverse
Lila-E8 trained on Fandom Multiverse (Monster + Time2026 + LMFDB + MathJax)
Model Details
- Architecture: Lila-E8 (Lie Lattice Attention)
- E8 Quantization: 240 roots
- Geometric Attention: E8-biased
- Training Steps: 9000
- Framework: PyTorch
Usage
import torch
from model.lila_e8 import LilaE8Config, LilaE8
# Load checkpoint
checkpoint = torch.load('checkpoint_step_9000.pt', map_location='cpu')
config = checkpoint['config']
model = LilaE8(config)
model.load_state_dict(checkpoint['model_state_dict'])
# Generate
# (see inference examples in repo)
Training
Trained on NVIDIA RTX 3080 Ti using Nix reproducible environment.
See sovereign-lila-e8 for training code.
- License
- agpl-3.0
- Created
- Mar 9, 2026
- Updated
- Mar 9, 2026