--- license: mit task_categories: - question-answering language: - en --- * `all_clean.json` * remove overlapped parts with ChEBI-20 test * remove no description SMILES * Format: ```python { SMILES : [ ["Please describe the molecule", DESCRIPTION], ..., ] } ``` * Stats * max tokens length: 6113 * min tokens length: 20 * mean tokens length: 191 * median tokens length: 149 * Total 326,689 single turn dialogue. Total 293,302 SMILES examples. * Size: \ Train: 264,391 \ Valid: 33,072 \ Test: 32,987 * conversation template ```python 'conversation':{ [ "from": "human", "value": , # randomly sample from query templates, ], [ "from": "gpt", "value": , # description about the given molecule ], } ``` * Each entry in the pickle contains: ``` "graph": [ "edge_index":, # array like (int64) "edge_feat":, # array like (int64) "node_feat":, # array like (int64) "num_nodes":, # int ], "conversation:: # see above ``` * 14 query templates ```python # bear me, we still use token instead of somewhat token { '\nCould you give me a brief overview of this molecule?', '\nCould you provide a description of this molecule?', '\nDescribe this molecule.', '\nPlease give me some details about this molecule.', '\nProvide a brief overview of this molecule.', '\nProvide a description of this molecule.', '\nWhat can you tell me about this molecule?', 'Could you give me a brief overview of this molecule?\n', 'Could you provide a description of this molecule?\n', 'Describe this molecule.\n', 'Please give me some details about this molecule.\n', 'Provide a brief overview of this molecule.\n', 'Provide a description of this molecule.\n', 'What can you tell me about this molecule?\n' } ```