divelab/DAPO_E2H-gsm8k-gaussian_0p25_0p75
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: divelab/DAPO_E2H-gsm8k-gaussian_0p25_0p75
Source: Original Platform
This commit is contained in:
ModelHub XC
2026-04-29 01:01:06 +08:00
commit f7ae09a148
15 changed files with 152150 additions and 0 deletions
Download Patch File Download Diff File Expand all files Collapse all files

10
.hydra/overrides.yaml Normal file
View File

@@ -0,0 +1,10 @@
- mode=train
- task=gsm8k
- algorithm=grpo
- algorithm.training.curriculum_schedule=gaussian
- model=qwen15
- algorithm.training.max_steps=1600
- algorithm.training.vllm_mode=colocate
- algorithm.training.scheduler_params.mu_exp=0.25
- algorithm.training.scheduler_params.sigma=0.75
- algorithm.training.scheduler_params.max_dapo_iter=4