Upload dataset

README.md

@@ -1310,6 +1310,84 @@ dataset_info:
     num_examples: 500
   download_size: 345419
   dataset_size: 815854
+- config_name: sample_metrics
+  features:
+  - name: model_id
+    dtype: string
+  - name: use_reranker
+    dtype: bool
+  - name: recall_at_1
+    dtype: float64
+  - name: recall_at_5
+    dtype: float64
+  - name: recall_at_10
+    dtype: float64
+  - name: by_outcome
+    struct:
+    - name: below_threshold
+      struct:
+      - name: count
+        dtype: int64
+      - name: recall_at_1
+        dtype: float64
+      - name: recall_at_10
+        dtype: float64
+      - name: recall_at_5
+        dtype: float64
+    - name: correct
+      struct:
+      - name: count
+        dtype: int64
+      - name: recall_at_1
+        dtype: float64
+      - name: recall_at_10
+        dtype: float64
+      - name: recall_at_5
+        dtype: float64
+    - name: not_retrieved
+      struct:
+      - name: count
+        dtype: int64
+      - name: recall_at_1
+        dtype: float64
+      - name: recall_at_10
+        dtype: float64
+      - name: recall_at_5
+        dtype: float64
+    - name: wrong_prediction
+      struct:
+      - name: count
+        dtype: int64
+      - name: recall_at_1
+        dtype: float64
+      - name: recall_at_10
+        dtype: float64
+      - name: recall_at_5
+        dtype: float64
+  - name: mrr
+    dtype: float64
+  - name: accuracy
+    dtype: float64
+  - name: correct
+    dtype: int64
+  - name: total
+    dtype: int64
+  - name: accuracy_ci_lower
+    dtype: float64
+  - name: accuracy_ci_upper
+    dtype: float64
+  - name: avg_latency_ms
+    dtype: float64
+  - name: embedding_dim
+    dtype: int64
+  - name: index_build_time_s
+    dtype: float64
+  splits:
+  - name: train
+    num_bytes: 2570
+    num_examples: 10
+  download_size: 15558
+  dataset_size: 2570
 - config_name: sample_traces
   features:
   - name: trace_id
@@ -1447,6 +1525,10 @@ configs:
   data_files:
   - split: train
     path: checkpoint_20260106_014829_intfloat_multilingual_e5_large_with_reranker_traces/train-*
+- config_name: sample_metrics
+  data_files:
+  - split: train
+    path: sample_metrics/train-*
 - config_name: sample_traces
   data_files:
   - split: train

sample_metrics/train-00000-of-00001.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:49f83fd0244e33dabb7780a6dbe52fad273323d07c5b5a96eb4206d712d3e44b
+size 15558