初始化项目，由ModelHub XC社区提供模型

Model: promotion/ronpo-qwen3-8b-fair-ronpo-full-expect-s42 Source: Original Platform
2026-07-18 11:54:12 +08:00
commit aac32a2b1a
18 changed files with 5506 additions and 0 deletions
--- a/README.md
+++ b/README.md
@@ -0,0 +1,12 @@
+---
+base_model: Qwen/Qwen3-8B
+library_name: transformers
+tags:
+- preference-optimization
+- ronpo
+- qwen3
+---
+
+# Qwen3-8B fair-demo checkpoint: ronpo_full_expect
+
+Validation-selected candidate: `ronpo_full_a`. Exact evaluator, selection, sweep, and training metadata are stored under `experiment/`.