Files
llama-sft-muon/train_results.json

8 lines
219 B
JSON
Raw Permalink Normal View History

{
"total_flos": 5.519394781646763e+19,
"train_loss": 3.57993624316737,
"train_runtime": 147576.5591,
"train_samples": 1444084,
"train_samples_per_second": 1.002,
"train_steps_per_second": 0.125
}