初始化项目，由ModelHub XC社区提供模型

Model: promotion/qwen3-8b-aaai27-flagship-ht-mnpo-helpfulness-s44 Source: Original Platform
2026-07-23 14:54:10 +08:00
commit 30c78df99f
20 changed files with 4373 additions and 0 deletions
--- a/train_results.json
+++ b/train_results.json
@@ -0,0 +1,9 @@
+{
+    "epoch": 0.8600609209819029,
+    "total_flos": 0.0,
+    "train_loss": 0.3132047750718064,
+    "train_runtime": 7069.0811,
+    "train_samples": 16743,
+    "train_samples_per_second": 2.037,
+    "train_steps_per_second": 0.127
+}