---
license: cc-by-4.0
language:
- en
tags:
- sentencepiece
- tokenizer
- fineweb
- parameter-golf
- bpe
size_categories:
- n<1K
---

# Parameter Golf SP16384 Tokenizer

SentencePiece BPE tokenizer (`vocab_size=16384`, `byte_fallback=True`) trained on the Parameter Golf 10B FineWeb corpus. Companion artifact to the chaoscontrol submission pipeline; produced to avoid re-running ~30–60 min of SP training on submission day.

## Files

- `fineweb_16384_bpe.model` — SentencePiece model (455 KB).
- `fineweb_16384_bpe.vocab` — Human-readable vocab sidecar (185 KB).

## Training configuration

| Setting | Value |
|---|---|
| `vocab_size` | 16384 |
| `model_type` | BPE |
| `byte_fallback` | True |
| `character_coverage` | 1.0 |
| `shuffle_input_sentence` | False (locks determinism) |
| `sp_seed` | 1337 |
| Training docs | 5,000,000 (first 5M post-val-split, per manifest convention) |
| Source corpus | [`willdepueoai/parameter-golf`](https://huggingface.co/datasets/willdepueoai/parameter-golf) → `docs_selected.jsonl` |
| Source revision | `9bb295ddab0e05d785b879661af7260fed5140fc` |
| SentencePiece | 0.2.1 |

Training was single-threaded in the BPE merge phase (SP 0.2.1 limitation; `num_threads` helps normalization only). Wall-clock on 28 vCPU Xeon 8480+ host: ~25 min SP training + ~30 min full-corpus tokenization.

## Reproducing from scratch

```bash
python scripts/build_sp_shards.py \
  --docs-path path/to/docs_selected.jsonl \
  --vocab-size 16384 \
  --sp-seed 1337 \
  --sp-train-docs 5000000 \
  --num-workers 28
```

Byte-identical outputs are guaranteed within a matching `(vocab_size, sp_seed, num_workers)` triple — SP's multi-threaded merge counting can drift on tie-breaks across thread counts. Use the same `--num-workers` for cross-machine determinism, or pin to `--num-workers 1` for strict identity.

## Use at tokenization time

```python
from huggingface_hub import hf_hub_download

model_path = hf_hub_download(
    repo_id="Natooka/parameter-golf-sp-tokenizers",
    filename="fineweb_16384_bpe.model",
    repo_type="dataset",
    local_dir="baselines/parameter_golf/tokenizers",
)
```

Then point `scripts/build_sp_shards.py --skip-train` at it to re-tokenize the corpus in ~5 min on 28 workers.

## License

CC-BY 4.0. Attribution appreciated for academic use. See upstream Parameter Golf competition for `docs_selected.jsonl` terms.