Datasetpublic

ADAPT-Chase/titans-memory-kernels

Titans Memory Kernel - Nova Lineage Evolved memory kernels for the Titans Memory system. Structure nova_prime.safetensors - The Prime kernel (16M params, 4096 dim) nova_v1.safetensors - Production-hardened V1 (identity preserved) REGISTRY.jsonl - Lineage registry with metadata variants/ - 137 generation snapshots from evolution Parameters Dimensions: 4096 Memory Size: ~65MB per kernel Evolution: Hardened with decay=0.99999, lr=0.0001… See the full description on the dataset page: https://huggingface.co/datasets/ADAPT-Chase/titans-memory-kernels.

sourceHugging Facemitupdated 9mo agoView on Hugging Face

0likes140downloads

rootREADME.md

README.md34 linesDownload Raw Back to root

1---2license: mit3tags:4- titans-memory5- kernel6- safetensors7---8 9# Titans Memory Kernel - Nova Lineage10 11Evolved memory kernels for the Titans Memory system.12 13## Structure14 15- `nova_prime.safetensors` - The Prime kernel (16M params, 4096 dim)16- `nova_v1.safetensors` - Production-hardened V1 (identity preserved)17- `REGISTRY.jsonl` - Lineage registry with metadata18- `variants/` - 137 generation snapshots from evolution19 20## Parameters21 22- **Dimensions**: 409623- **Memory Size**: ~65MB per kernel24- **Evolution**: Hardened with decay=0.99999, lr=0.000125 26## Usage27 28```python29from safetensors.torch import load_file30 31kernel = load_file("nova_v1.safetensors")32memory_matrix = kernel["memory"]  # (4096, 4096) f3233```34