diff --git a/src/training/plot_metrics.py b/src/training/plot_metrics.py
index df97c62..c8db1aa 100644
--- a/src/training/plot_metrics.py
+++ b/src/training/plot_metrics.py
@@ -25,26 +25,54 @@ sns.set_theme(style="whitegrid", palette="deep", font_scale=1.1)
 
 
 def plot_train_metrics(train_df: pd.DataFrame, out_dir: Path) -> None:
-    """Plot training metrics: loss, temperature, gates, lr."""
-    fig, axes = plt.subplots(2, 2, figsize=(14, 10))
+    """Plot training metrics: loss, recall, AP, temperature, gates, lr."""
+    fig, axes = plt.subplots(2, 3, figsize=(20, 10))
     fig.suptitle("Training Metrics", fontsize=16, fontweight="bold")
 
     # 1. Loss.
     ax = axes[0, 0]
-    sns.lineplot(data=train_df, x="epoch", y="total", ax=ax, marker="o", linewidth=2)
-    ax.set_title("Loss")
+    col = "train_loss" if "train_loss" in train_df.columns else "total"
+    if col in train_df.columns:
+        sns.lineplot(data=train_df, x="epoch", y=col, ax=ax, marker="o", linewidth=2)
+    ax.set_title("Train Loss")
     ax.set_xlabel("Epoch")
     ax.set_ylabel("InfoNCE Loss")
 
-    # 2. Temperature (tau).
+    # 2. Recall@K (train).
     ax = axes[0, 1]
-    sns.lineplot(data=train_df, x="epoch", y="temperature", ax=ax, marker="s", linewidth=2, color="orange")
+    for k in [1, 5, 10]:
+        col = f"r@{k}_q2g"
+        if col in train_df.columns:
+            df_valid = train_df.dropna(subset=[col])
+            if not df_valid.empty:
+                sns.lineplot(data=df_valid, x="epoch", y=col, ax=ax, marker="o", linewidth=2, label=f"R@{k}")
+    ax.set_title("Train Recall@K (drone → sat)")
+    ax.set_xlabel("Epoch")
+    ax.set_ylabel("Recall")
+    ax.set_ylim(0, 1)
+    ax.legend()
+
+    # 3. Average Precision (train).
+    ax = axes[0, 2]
+    if "ap_q2g" in train_df.columns:
+        df_valid = train_df.dropna(subset=["ap_q2g"])
+        if not df_valid.empty:
+            sns.lineplot(data=df_valid, x="epoch", y="ap_q2g", ax=ax, marker="s", linewidth=2, color="purple")
+    ax.set_title("Train AP (drone → sat)")
+    ax.set_xlabel("Epoch")
+    ax.set_ylabel("AP")
+    ax.set_ylim(0, 1)
+
+    # 4. Temperature (tau).
+    ax = axes[1, 0]
+    if "temperature" in train_df.columns:
+        sns.lineplot(data=train_df, x="epoch", y="temperature", ax=ax, marker="s", linewidth=2, color="orange")
     ax.set_title("Temperature (τ)")
     ax.set_xlabel("Epoch")
     ax.set_ylabel("τ")
 
-    # 3. Gate values.
-    ax = axes[1, 0]
+    # 5. Gate values.
+    ax = axes[1, 1]
     if "gate_q" in train_df.columns and "gate_g" in train_df.columns:
         sns.lineplot(data=train_df, x="epoch", y="gate_q", ax=ax, marker="o", linewidth=2, label="gate_q (drone)")
         sns.lineplot(data=train_df, x="epoch", y="gate_g", ax=ax, marker="s", linewidth=2, label="gate_g (sat)")
@@ -54,9 +82,10 @@ def plot_train_metrics(train_df: pd.DataFrame, out_dir: Path) -> None:
     ax.set_ylabel("Image weight")
     ax.set_ylim(0, 1)
 
-    # 4. Learning rate.
-    ax = axes[1, 1]
-    sns.lineplot(data=train_df, x="epoch", y="lr", ax=ax, marker="^", linewidth=2, color="green")
+    # 6. Learning rate.
+    ax = axes[1, 2]
+    if "lr" in train_df.columns:
+        sns.lineplot(data=train_df, x="epoch", y="lr", ax=ax, marker="^", linewidth=2, color="green")
     ax.set_title("Learning Rate")
     ax.set_xlabel("Epoch")
     ax.set_ylabel("LR")