初始化项目，由ModelHub XC社区提供模型

Model: SeongryongJung/qwen3-4b-tooluse-rlsd-ema005 Source: Original Platform
2026-09-21 02:04:18 +08:00
commit 5323ac29f2
17 changed files with 152421 additions and 0 deletions
--- a/metrics.json
+++ b/metrics.json
@@ -0,0 +1,51 @@
+{
+  "wandb_run": "run-20260630_063402-76k27fnx",
+  "wandb_output_log": "/mnt/mole/SDPO/L2T/wandb/run-20260630_063402-76k27fnx/files/output.log",
+  "metric_key": "val-aux/tooluse/reward/mean@16",
+  "evals": [
+    {
+      "step": 10,
+      "mean16": 0.5992647058823529
+    },
+    {
+      "step": 20,
+      "mean16": 0.5606617647058824
+    },
+    {
+      "step": 30,
+      "mean16": 0.6029411764705882
+    },
+    {
+      "step": 40,
+      "mean16": 0.5882352941176471
+    },
+    {
+      "step": 50,
+      "mean16": 0.5955882352941176
+    },
+    {
+      "step": 60,
+      "mean16": 0.609375
+    },
+    {
+      "step": 70,
+      "mean16": 0.6341911764705882
+    },
+    {
+      "step": 80,
+      "mean16": 0.6121323529411765
+    },
+    {
+      "step": 90,
+      "mean16": 0.5845588235294118
+    },
+    {
+      "step": 100,
+      "mean16": 0.6011029411764706
+    }
+  ],
+  "best_step": 70,
+  "best_mean16": 0.6341911764705882,
+  "final_step": 100,
+  "final_mean16": 0.6011029411764706
+}