open-thoughts/OpenThinkerAgent-8B-ColdStartSFTForRL
1
0
Fork 0
Code Issues Pull Requests Actions Projects Releases Wiki Activity

初始化项目，由ModelHub XC社区提供模型

Browse Source 
Model: open-thoughts/OpenThinkerAgent-8B-ColdStartSFTForRL
Source: Original Platform
This commit is contained in:
ModelHub XC
2026-07-26 15:18:45 +08:00
commit 4ebbc76875
23 changed files with 162423 additions and 0 deletions
Download Patch File Download Diff File Expand all files Collapse all files

16
train_results.json Normal file
View File

@@ -0,0 +1,16 @@
{
"achieved_tflops_per_gpu": 5.143096781650618,
"achieved_tflops_per_gpu_theoretical": 216.82589202068274,
"epoch": 7.0,
"loss_nan_ranks": 0,
"loss_rank_avg": 0.07686486095190048,
"mfu_percent": 0.5200300082558764,
"mfu_percent_theoretical": 21.923750457096332,
"total_flos": 2.65446196155043e+18,
"train_loss": 0.30306281665335555,
"train_runtime": 64515.1665,
"train_samples_per_second": 1.024,
"train_steps_per_second": 0.064,
"valid_targets_mean": 6255.1,
"valid_targets_min": 3251
}