GQA-ru

GQA-ru This is a translated version of original GQA dataset and stored in format supported for lmms-eval pipeline. For this dataset, we: Translate the original one with gpt-4-turbo Filter out unsuccessful translations, i.e. where the model protection was triggered Manually validate most common errors Dataset Structure Dataset includes both train and test splits translated from original train balanced and testdev balanced. Train split includes 27519 images with… See the full description on the dataset page:

Source: Hugging Face Hub (deepvk/GQA-ru). Metadata imported from the dataset’s Hub tags.