Files
Qwen1.5-MOE-sft-ESFT-summary/all_results.json

8 lines
217 B
JSON
Raw Permalink Normal View History

{
"total_flos": 1.770321796592042e+18,
"train_loss": 1.3923614178746209,
"train_runtime": 2541.2937,
"train_samples": 19587,
"train_samples_per_second": 7.707,
"train_steps_per_second": 0.241
}