初始化项目，由ModelHub XC社区提供模型

Model: promotion/qwen3-8b-aaai27-flagship-ht-mnpo-helpfulness-s42 Source: Original Platform
2026-07-23 13:58:09 +08:00
commit 0738f2ca0a
20 changed files with 5093 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": 0.3323093075222439,
+    "train_runtime": 7030.1668,
+    "train_samples": 16743,
+    "train_samples_per_second": 2.048,
+    "train_steps_per_second": 0.128
+}