W-61/llama-3-8b-base-margin-dpo-ultrafeedback-8xh200
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: W-61/llama-3-8b-base-margin-dpo-ultrafeedback-8xh200
Source: Original Platform
This commit is contained in:
ModelHub XC
2026-04-24 11:32:05 +08:00
commit 15c4f36648
498 changed files with 6803 additions and 0 deletions
Download Patch File Download Diff File Expand all files Collapse all files

16
eval_results.json Normal file
View File

@@ -0,0 +1,16 @@
{
"epoch": 0.9989528795811519,
"eval_logits/chosen": -0.843854546546936,
"eval_logits/rejected": -0.8338058590888977,
"eval_logps/chosen": -410.1976013183594,
"eval_logps/ref_chosen": -280.7076110839844,
"eval_logps/ref_rejected": -280.1632385253906,
"eval_logps/rejected": -484.2484130859375,
"eval_loss": 0.5346022844314575,
"eval_margin_dpo/margin_mean": 74.59521484375,
"eval_margin_dpo/margin_std": 98.45130157470703,
"eval_runtime": 50.8318,
"eval_samples": 2000,
"eval_samples_per_second": 39.345,
"eval_steps_per_second": 0.63
}