JaSpa99 d3d58a37e5 Bert: use Tensor.scaled_dot_product_attention (#1528) ...

* use scaled attn from Tensor

* add a test for bert

* linter

* no more tokenizer

* without loading weights

* remove prints

* tribute to linter lords

* smaller input and less runs

* small bert

2023-08-12 08:46:04 -07:00

2.1 KiB

Raw Blame History

View Raw