Files
sft__stackexchange-tezos-sa…/train_results.json

16 lines
538 B
JSON
Raw Permalink Normal View History

{
"achieved_tflops_per_gpu": 5.062071165920845,
"achieved_tflops_per_gpu_theoretical": 237.15331939762177,
"epoch": 7.0,
"loss_nan_ranks": 0,
"loss_rank_avg": 0.13302914798259735,
"mfu_percent": 0.3577435452947593,
"mfu_percent_theoretical": 16.759951900892,
"total_flos": 1.4527566279155384e+18,
"train_loss": 0.4386636906199985,
"train_runtime": 8968.3932,
"train_samples_per_second": 6.727,
"train_steps_per_second": 0.07,
"valid_targets_mean": 3211.4,
"valid_targets_min": 220
}