Views
No views yet
| Architecture | Transformer (RMSNorm, RoPE, QK-norm, ReLU², logit softcapping, residual scalars) |
| Parameters | 25.2M |
| Depth | 8 (n_embd=512, n_layer=8, n_head=8) |
| Vocab size | 186 (WordLevel, 2-digit pair encoding) |
| Training data | 174K OC20 structures |
| Training time | 97 min on TPU v6e-8 |
| Framework | JAX / Flax |
| Metric | Result |
|---|---|
| Parseable | 96/100 |
| Element Match | 96/100 |
| Generation Validity | 96/100 (96.0%) |
| Uniqueness | 96/96 (100.0%) |
| Novelty | 96/96 (100.0%) |
| Min Distance (≥ 0.5Å) | 83/96 (86.5%) |
1from catalyst.hub import download_checkpoint
2from catalyst.config import CatalystConfig
3from catalyst.generate import generate_samples
4
5# Download checkpoint
6ckpt_path = download_checkpoint("everythingchalna/nanocatalyst")
7config = CatalystConfig.load(ckpt_path / "config.json")
8
9# Load params and generate (see README for full example)config.json — Model configurationparams/ — Orbax checkpoint (model parameters)tokenizer.json — HuggingFace WordLevel tokenizertokenizer_stats.json — Tokenizer coverage statistics1@software{nanocatalyst,
2 title = {nanocatalyst},
3 url = {https://github.com/everythingchalna/nanocatalyst},
4 license = {MIT}
5}