初始化项目，由ModelHub XC社区提供模型

Model: Gueule-d-ange/aup-fullft-kto_w1_mmd-w1lam8.4e-4_mmdrho8.4e-4_kr0.1-seed3 Source: Original Platform
2026-07-25 01:58:10 +08:00
commit e90962e0d2
20 changed files with 253745 additions and 0 deletions
--- a/all_results.json
+++ b/all_results.json
@@ -0,0 +1,8 @@
+{
+    "epoch": 3.0,
+    "total_flos": 9.960196480656998e+17,
+    "train_loss": 0.5863801720640615,
+    "train_runtime": 11092.9859,
+    "train_samples_per_second": 13.361,
+    "train_steps_per_second": 0.209
+}