BIG Bench but it doesn't require the hellish dependencies (tensorflow, pypi bigbench, protobuf) of the official version. Code to reproduce: https://colab.research.google.com/drive/1MKdLdF7oqrSQCeavAcsEnPdI85kD0LzU?usp=sharing Datasets are capped to 50k examples to keep things light. I also removed the default split when train was available also to save space, as default=train+val.
We use cookies for essential functionality and analytics. You can accept or reject analytics cookies.Cookie policy