Model: laion/open-thoughts-4-code-qwen3-32b-annotated-7k_qwen3-8B_8k Source: Original Platform
16 lines
541 B
JSON
16 lines
541 B
JSON
{
|
|
"achieved_tflops_per_gpu": 42.42053490110656,
|
|
"achieved_tflops_per_gpu_theoretical": 4844.70352483713,
|
|
"epoch": 1.0,
|
|
"loss_nan_ranks": 0,
|
|
"loss_rank_avg": 0.19463524222373962,
|
|
"mfu_percent": 4.289235075946063,
|
|
"mfu_percent_theoretical": 489.8587992757462,
|
|
"total_flos": 1.4231344961965523e+19,
|
|
"train_loss": 0.06952266532071043,
|
|
"train_runtime": 41935.3062,
|
|
"train_samples_per_second": 22.874,
|
|
"train_steps_per_second": 1.43,
|
|
"valid_targets_mean": 7497.0,
|
|
"valid_targets_min": 7481
|
|
} |