| {"$id":"https://huggingface.co/datasets/cqqq/GSXray/raw/main/schema.json","$schema":"https://json-schema.org/draft/2020-12/schema","additionalProperties":false,"legacy_array_file":"annotations/sft_merged_44k.json omits id but otherwise uses conversations/images exactly as above","properties":{"conversations":{"maxItems":2,"minItems":2,"prefixItems":[{"additionalProperties":false,"properties":{"from":{"const":"human"},"value":{"type":"string"}},"required":["from","value"],"type":"object"},{"additionalProperties":false,"properties":{"from":{"const":"gpt"},"value":{"type":"string"}},"required":["from","value"],"type":"object"}],"type":"array"},"id":{"pattern":"^[0-9]{6}$","type":"string"},"images":{"maxItems":2,"minItems":2,"prefixItems":[{"pattern":"^images/top/[0-9]{6}\\.jpg$","type":"string"},{"pattern":"^images/side/[0-9]{6}\\.jpg$","type":"string"}],"type":"array"}},"required":["id","conversations","images"],"title":"GSXray train.jsonl record","type":"object"} | |