Model: open-sci/sft__ot30k_Qwen3-1.7B-Base-SFT-Tulu3-decontaminated Source: Original Platform
12 lines
336 B
JSON
12 lines
336 B
JSON
{
|
|
"epoch": 5.0,
|
|
"loss_nan_ranks": 0,
|
|
"loss_rank_avg": 0.4691426753997803,
|
|
"total_flos": 1658037051588608.0,
|
|
"train_loss": 1.0214215231449046,
|
|
"train_runtime": 6105.894,
|
|
"train_samples_per_second": 24.566,
|
|
"train_steps_per_second": 0.192,
|
|
"valid_targets_mean": 13499.5,
|
|
"valid_targets_min": 3345
|
|
} |