BERT, from scratch
A 7.5M-parameter Bengali BERT, pre-trained from nothing on a laptop, that scores above published mBERT and IndicBERT on Bengali news classification.
বিশ্বমানব হবি যদিশাশ্বত বাঙ্গালী হ’॥
A 7.5M-parameter Bengali BERT, pre-trained from nothing on a laptop, that scores above published mBERT and IndicBERT on Bengali news classification.
The original transformer rebuilt component by component, then trained to translate English into Bengali on one laptop.
A ViT trained on 1,935 photographs of Bengali terracotta temples — a dataset I had to build first — to test the paper's claim about pre-training. It failed, exactly as predicted.