Step 2 of 4 · ProjectsModels
hscode-classifier-en (ENTUM-AI)
Deep-learning HS code classifier for English product descriptions, with weights.
What it does
A PyTorch model on bert-base-uncased that predicts Harmonised System codes at 2-, 4- and 6-digit level from an English product description, giving five candidates and confidence per level. It needs torch and transformers and the custom class in the bundled inference.py. The card says English only, coverage limited to codes in the training data, and scores are rankings, not probabilities; MIT.
Where it fits on the map
Similar projects
Of the same kind, for a shared jurisdiction or tax area. Not a ranking.
- gemma4-hscode-classifierLoRA adapter classifying product descriptions to 8-digit HS codes with Cambodian trade rates; adapter only.
- hs-code-classifier-en (Kenpache)HS6 classifier fine-tuned from bge-m3 (71.9% top-1 on full descriptions in its test set), with weights.
- hscodeclassifyModernBERT classifier of product descriptions into the 20 HS chapters; chapter level only.
- UK Trade Tariff FPO commodity searchMachine-learning service from the UK tariff team that suggests commodity codes from parcel descriptions.
- AI-Powered-Tax-OptimizationRandom Forest model and Flask API that return one optimised tax figure from nine financial inputs.Inactive since 2024Demo
- hs-code-classifier (ShahinHasanov90)TF-IDF and LightGBM HS code classifier with a FastAPI service for English, Russian and Azerbaijani text.Demo
Maintain this project?
Link to this page from your README, so readers find the facts, the notes and where it fits. The badge is a small SVG served from this site.
[](https://tax-oss.com/projects/hscode-classifier-en-entum/)