Tomáš Houfek

icon company @BBMRI-cz icon location Brno, Czech Republic icon location Student of Artificial Intelligence at FI MUNI Currently working as DevOps, SW developer and AI Researcher at MMCI.

Results 2 comments of Tomáš Houfek

Text attribution fails for XLM-Roberta models

I've managed to solve the problem it seems. The problem is that xlm-roberta-like models have `(token_type_embeddings): Embedding(1, 1024)` Meaning token_type_ids vector can be of max_length = 1024 and only consist...

Choice pre-selection does not work if `perRegion` is set

Hi, seems this is still happening.