feat(benchmark): support multi-dataset evaluation with configurable top-k list

- Evaluate multiple datasets in a single run (CIFAR10 + CIFAR100)
- Report Recall@K for a list of K values from one underlying search
- Save results to disk: summary.md, metrics.csv, per-dataset predictions.csv, confusion_matrix.csv
- Richer evaluation output: query_labels, topk_ids, topk_labels for downstream analysis
- Add --dataset and --top-k CLI overrides for benchmark command
- Update config schema: dataset→datasets, top_k→top_k_list

mini-nav/configs/config.yaml

@@ -22,14 +22,18 @@ dataset:
     seed: 42
 
 benchmark:
-    dataset:
-        source_type: "huggingface"
-        path: "uoft-cs/cifar100"
-        img_column: "img"
-        label_column: "fine_label"
+    datasets:
+        - source_type: "huggingface"
+          path: "uoft-cs/cifar100"
+          img_column: "img"
+          label_column: "fine_label"
+        - source_type: "huggingface"
+          path: "uoft-cs/cifar10"
+          img_column: "img"
+          label_column: "label"
     task:
         name: "recall_at_k"
         type: "retrieval"
-        top_k: 1
+        top_k_list: [1, 5, 10]
     batch_size: 64
     model_table_prefix: "benchmark"