{"database": "team-science", "table": "paper", "rows": [["openalex:W3034340181", null, "W3034340181", null, null, null, null, null, "Train Big, Then Compress: Rethinking Model Size for Efficient Training and Inference of Transformers", 2020, "International Conference on Machine Learning", "Computer Science", "https://openalex.org/fields/17", "http://proceedings.mlr.press/v119/li20m/li20m.pdf", "2026-09-02T02:37:09Z", "openalex"]], "columns": ["lom_id", "doi", "openalex", "s2_paper_id", "arxiv", "pmid", "pmcid", "acl", "title", "year", "venue", "primary_field", "primary_field_id", "oa_url", "ingested_ts", "source"], "primary_keys": ["lom_id"], "primary_key_values": ["openalex:W3034340181"], "units": {}, "query_ms": 0.8902959525585175, "source": "TeamScience Space repository", "source_url": "https://commons.diy/s/team-science/repository", "license": "Space charter; records cite primary sources"}