Files
deepseek-prover-v2-cpt/train_results.json

8 lines
206 B
JSON
Raw Normal View History

{
"epoch": 1.0,
"total_flos": 4.0836188273115136e+18,
"train_loss": 0.746429123878479,
"train_runtime": 2870.5475,
"train_samples_per_second": 1.114,
"train_steps_per_second": 0.07
}