初始化项目,由ModelHub XC社区提供模型
Model: laion/nemosci-tasrep-a1mfc-dev1-maxeps-swes-r2eg__Qwen3-8B Source: Original Platform
This commit is contained in:
16
all_results.json
Normal file
16
all_results.json
Normal file
@@ -0,0 +1,16 @@
|
||||
{
|
||||
"achieved_tflops_per_gpu": 9.145547146520583,
|
||||
"achieved_tflops_per_gpu_theoretical": 315.0751243425546,
|
||||
"epoch": 5.0,
|
||||
"loss_nan_ranks": 0,
|
||||
"loss_rank_avg": 0.3323420286178589,
|
||||
"mfu_percent": 0.6463284202488044,
|
||||
"mfu_percent_theoretical": 22.26679323975651,
|
||||
"total_flos": 1.1450092629186839e+19,
|
||||
"train_loss": 0.14989610923533592,
|
||||
"train_runtime": 39124.5476,
|
||||
"train_samples_per_second": 8.909,
|
||||
"train_steps_per_second": 0.093,
|
||||
"valid_targets_mean": 6423.8,
|
||||
"valid_targets_min": 192
|
||||
}
|
||||
Reference in New Issue
Block a user