初始化项目,由ModelHub XC社区提供模型
Model: DCAgent/g1_min_episodes_e1_gpt_long_2x_tacc-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": 0.003643509490452907,
|
||||
"achieved_tflops_per_gpu_theoretical": 443.815497729835,
|
||||
"epoch": 7.0,
|
||||
"loss_nan_ranks": 0,
|
||||
"loss_rank_avg": 0.2079296112060547,
|
||||
"mfu_percent": 0.00025749183678112415,
|
||||
"mfu_percent_theoretical": 31.365052843097878,
|
||||
"total_flos": 4008973724024832.0,
|
||||
"train_loss": 0.3080715393729617,
|
||||
"train_runtime": 45846.0647,
|
||||
"train_samples_per_second": 3.134,
|
||||
"train_steps_per_second": 0.131,
|
||||
"valid_targets_mean": 5229.1,
|
||||
"valid_targets_min": 2750
|
||||
}
|
||||
Reference in New Issue
Block a user