Files
Qwen3-8B_exp-swd-swesmith-w…/all_results.json

16 lines
545 B
JSON
Raw Normal View History

{
"achieved_tflops_per_gpu": 0.0020004562843468124,
"achieved_tflops_per_gpu_theoretical": 189.7226555180974,
"epoch": 8.0,
"loss_nan_ranks": 0,
"loss_rank_avg": 0.012045512907207012,
"mfu_percent": 0.0006411718860085937,
"mfu_percent_theoretical": 60.80854343528763,
"total_flos": 2887866713112576.0,
"train_loss": 0.1099344994546549,
"train_runtime": 45112.6253,
"train_samples_per_second": 1.778,
"train_steps_per_second": 0.056,
"valid_targets_mean": 4840.4,
"valid_targets_min": 726
}