Instructions to use vpermilp/spellman with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- fastText
How to use vpermilp/spellman with fastText:
from huggingface_hub import hf_hub_download import fasttext model = fasttext.load_model(hf_hub_download("vpermilp/spellman", "model.bin")) - Notebooks
- Google Colab
- Kaggle
Commit History
17 days ago
-
v16: label-noise cleanup retrain (theta 0.67), int8-col
Verified signature539a1b1
20 days ago
-
One storage format: int8 with per-column scales at the root; drop f16/int8-row/int8-col/fp8 variants
Verified signature1d5e653
-
card: human-first rewrite
Verified signature22f386e -
card: v14 measured throughput
Verified signature933d44e -
v14: 2^18 buckets, k=512 truncation, short-floor 0.5, theta by error-F1; int8-row root
Verified signatured0a7217
-
v13c: crawl datasets (vpermilp/lid-*) + 120k cap; f16 root artifact
Verified signature5790e63
-
v12 card: GlotLID same-split baseline (90.47)
Verified signature5e060eb -
v12: commercial-clean rebuild (NC-free mix, orthographic gates, scaled hard negatives)
Verified signature341a94e -
docs: v11c card fixes — tatoeba ladder, 12 diverse lanes (algo=6), literary lane, 32k cap, dataset link (was: stale data_mix4 pointer)
Verified signature5cb3fe4
-
Upload README.md with huggingface_hub
Verified signaturead0304c -
Upload README.md with huggingface_hub
Verified signatureae27fb6
-
Upload README.md with huggingface_hub
Verified signaturea66fbd3 -
Upload README.md with huggingface_hub
Verified signature811e8d3
-
Upload README.md with huggingface_hub
Verified signature04bce66 -
Upload README.md with huggingface_hub
Verified signature20aed59 -
Upload README.md with huggingface_hub
Verified signaturef16b2bf