add DocVQA scoring

llama_stack/templates/open-benchmark/run.yaml

@@ -188,6 +188,22 @@ datasets:
     uri: huggingface://datasets/llamastack/bfcl_v3?split=train
   metadata: {}
   dataset_id: bfcl
+<<<<<<< HEAD
+=======
+  provider_id: huggingface
+- dataset_schema:
+    chat_completion_input:
+      type: string
+    expected_answer:
+      type: string
+  url:
+    uri: https://huggingface.co/datasets/llamastack/docvqa
+  metadata:
+    path: llamastack/docvqa
+    split: train
+  dataset_id: docvqa
+  provider_id: huggingface
+>>>>>>> 81df5a4e (add DocVQA scoring)
 scoring_fns: []
 benchmarks:
 - dataset_id: simpleqa
@@ -215,6 +231,11 @@ benchmarks:
   - basic::bfcl
   metadata: {}
   benchmark_id: meta-reference-bfcl
+- dataset_id: docvqa
+  scoring_functions:
+  - basic::docvqa
+  metadata: {}
+  benchmark_id: meta-reference-docvqa
 tool_groups:
 - toolgroup_id: builtin::websearch
   provider_id: tavily-search