Mascerade/supervised-product-matching
⚖️ Neural network for product matching, aka classifying whether two product titles represent the same entity
Built on CharacterBERT with optional stacked Transformer layers, the model handles character-level semantics to match electronics product titles despite variations in specification ordering and descriptive verbosity. Training leverages PyTorch with multiple architecture variants (concatenation-based embeddings, custom Transformers) and integrates HuggingFace Transformers for BERT components. Includes web scraping utilities for data collection and a modular package design enabling portability across downstream applications.
No commits in the last 6 months.
Stars
67
Forks
16
Language
Python
License
MIT
Category
Last pushed
May 28, 2023
Commits (30d)
0
Get this data via API
curl "https://pt-edge.onrender.com/api/v1/quality/nlp/Mascerade/supervised-product-matching"
Open to everyone — 100 requests/day, no key needed. Get a free key for 1,000/day.
Higher-rated alternatives
jenojp/negspacy
spaCy pipeline object for negating concepts in text
dmlls/negate
negate_sentence(A Python module that doesn't negate sentences.)
sunzeyeah/item-alignment
ccks2022 task9 subtask2 商品同款识别
sugatagh/E-commerce-Text-Classification
Proper categorization of e-commerce products enhances the user experience and achieves better...
MaggieMeow/neko
Product categorization using sequence-to-sequence models and others.