初始化项目,由ModelHub XC社区提供模型

Model: DCAgent/FourDatasetMixQwen3_8B
Source: Original Platform
This commit is contained in:
ModelHub XC
2026-05-13 04:30:37 +08:00
commit 2a5b291815
22 changed files with 159959 additions and 0 deletions

16
train_results.json Normal file
View File

@@ -0,0 +1,16 @@
{
"achieved_tflops_per_gpu": 4.132335436646405,
"achieved_tflops_per_gpu_theoretical": 111.80085057316455,
"epoch": 5.0,
"loss_nan_ranks": 0,
"loss_rank_avg": 0.01754429191350937,
"mfu_percent": 1.324466486104617,
"mfu_percent_theoretical": 35.833605952937354,
"total_flos": 1.586502852660953e+18,
"train_loss": 0.19733358554124833,
"train_runtime": 95981.0062,
"train_samples_per_second": 0.52,
"train_steps_per_second": 0.033,
"valid_targets_mean": 6467.2,
"valid_targets_min": 3371
}