huggingface__transformers-13573

Upcasting of attention computation for reliable pretraining of GPT-2 models

Verdict: no gold-patch lines found in the issue text (rule check only, no model read).

No gold-patch lines found in the issue text (rule check only, no model read). Upstream fix commit 3a8de58c51 dated 2021-10-04.

All SWE-PolyBench (full set sample) tasks · more huggingface/transformers tasks · How the check works · Report a correction