AMBERT: BERT with Multi-Grained Tokenization Achieves SOTA Results on English and Chinese NLU Tasks | Synced

AMBERT (A Multigrained BERT) leverages both fine-grained and coarse-grained tokenizations to achieve SOTA performance on English and Chinese language tasks.

By · · 1 min read

Source: Synced | AI Technology & Industry Review

AMBERT (A Multigrained BERT) leverages both fine-grained and coarse-grained tokenizations to achieve SOTA performance on English and Chinese language tasks.