thinkingdbx/codebert-permissive-embed

1{2  "stage": "1 of 2: masked language model pretraining, from scratch",3  "documents_kept": 1980241,4  "documents_seen": 2047089,5  "permissive_fraction": 0.9673,6  "licences_kept": 171,7  "licences_excluded": 139,8  "unlicensed_kept": 0,9  "pii_redacted": 1384479,10  "tokens_trained": 3973189632,11  "steps": 30402,12  "final_loss": 4.3267,13  "vocab_size": 32000,14  "parameters": 11070000015}