{"display_name": "WebLINX", "total_episodes": 969, "total_steps": 24417, "data_format": "hdf5", "observation_space": "{\"type\": \"Dict\", \"subspaces\": {\"action_history\": {\"type\": \"Text\", \"max_length\": 16384, \"min_length\": 1, \"charset\": \"0123456789ABCDEFGHIJKLMNOPQRSTUVWXYZabcdefghijklmnopqrstuvwxyz\"}, \"candidates\": {\"type\": \"Text\", \"max_length\": 16384, \"min_length\": 1, \"charset\": \"0123456789ABCDEFGHIJKLMNOPQRSTUVWXYZabcdefghijklmnopqrstuvwxyz\"}, \"clean_html\": {\"type\": \"Text\", \"max_length\": 16384, \"min_length\": 1, \"charset\": \"0123456789ABCDEFGHIJKLMNOPQRSTUVWXYZabcdefghijklmnopqrstuvwxyz\"}, \"utterances\": {\"type\": \"Text\", \"max_length\": 8192, \"min_length\": 1, \"charset\": \"0123456789ABCDEFGHIJKLMNOPQRSTUVWXYZabcdefghijklmnopqrstuvwxyz\"}, \"viewport\": {\"type\": \"Text\", \"max_length\": 1024, \"min_length\": 1, \"charset\": \"0123456789ABCDEFGHIJKLMNOPQRSTUVWXYZabcdefghijklmnopqrstuvwxyz\"}}}", "action_space": "{\"type\": \"Text\", \"max_length\": 1024, \"min_length\": 1, \"charset\": \"0123456789ABCDEFGHIJKLMNOPQRSTUVWXYZabcdefghijklmnopqrstuvwxyz\"}", "dataset_id": "webagents/weblinx-v0", "code_permalink": "https://github.com/Farama-Foundation/minari-dataset-generation-scripts", "author": ["Zden\u011bk Kasner", "Siva Reddy", "Xing Han L\u00f9"], "author_email": ["xing.han.lu@mail.mcgill.ca"], "minari_version": "0.5.2", "dataset_size": 135.1, "description": "This dataset is a Minari formatted version of [WebLINX](https://huggingface.co/datasets/McGill-NLP/WebLINX). WebLINX is a benchmark for evaluating web navigation agents, and provides a dataset featuring complex goal-oriented tasks within interactive web environments."}