Upload folder using huggingface_hub

README.md

@@ -0,0 +1,36 @@
+---
+license: other
+task_categories:
+- visual-question-answering
+tags:
+- benchability
+- figure4
+---
+
+# BenchAbility Figure 4 -- held-out eval
+
+The 10% candidate-dev half of the same 884,143-row pool the two training mixtures are drawn from,
+balanced per capability. Split **by image**, so no picture here appears in either mixture, and both
+arms are equally blind to it.
+
+| capability | n |
+|---|---:|
+| chart_reading / chart_reasoning | 250 / 250 |
+| table_lookup / table_reasoning | 250 / 250 |
+| document_qa / document_text_reading | 250 / 250 |
+| diagram_and_infographic_understanding | 250 |
+| scene_text_recognition | 250 |
+| key_information_extraction | 168 (all that exists) |
+
+Balanced rather than proportional: the pool is 38% `chart_reasoning`, so a proportional dev set would
+measure that leaf precisely and the scarce ones not at all -- and the scarce ones are where the two
+arms differ most. One row per image, because a second question on the same picture is not an
+independent measurement.
+
+**This is an in-domain dev set, not the paper's evaluation.** Same 20 sources, held out by image. It
+answers "did training move this capability at all", which separates a broken run from a real one. It
+cannot separate "learned the capability" from "learned these 20 datasets" -- that needs the
+unseen-source set (ChartQAPro, OCRBench v2, DUDE, MME-RealWorld) which is not in this pool.
+
+Score with `evaluate.py` from the training bundle; measure the base model first, then each
+checkpoint, and report `gain = checkpoint - base` per capability.

data/train-00000-of-00003.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:e5ce0673b65dfe81d8023dc190284e380f8ff339d0f94c03481e97bfdd0815bc
+size 140862214

data/train-00001-of-00003.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:29827997edc88a329aea76b3539fbb4dff523b8a8077a342ad6d1292c327ea0d
+size 212588753

data/train-00002-of-00003.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:d0c2f9c4322794f18e531774a7e78f7463b3a5c26304c3aab06a5f3b4c98b9a4
+size 148329280

eval_manifest.json

@@ -0,0 +1,161 @@
+{
+  "spec_fingerprint": "37a89086c195",
+  "split_salt": "fig4-candidate-split-v1",
+  "pick_salt": "fig4-eval-pick-v1",
+  "per_capability": 250,
+  "source_cap": 0.5,
+  "note": "held-out candidate-dev, in-domain. NOT the paper's cross-source eval set.",
+  "per_capability_detail": {
+    "chart_reasoning": {
+      "n": 250,
+      "available": 50612,
+      "sources": {
+        "figureqa": 115,
+        "plotqa": 80,
+        "dvqa": 51,
+        "chartqa": 3,
+        "infovqa": 1
+      }
+    },
+    "table_reasoning": {
+      "n": 250,
+      "available": 3401,
+      "sources": {
+        "tabmwp": 116,
+        "multihiertt": 62,
+        "tatqa": 45,
+        "hitab": 24,
+        "scienceqa_image": 2,
+        "docvqa": 1
+      }
+    },
+    "diagram_and_infographic_understanding": {
+      "n": 250,
+      "available": 775,
+      "sources": {
+        "ai2d": 123,
+        "infovqa": 79,
+        "scienceqa_image": 43,
+        "iconqa": 2,
+        "visualmrc": 2,
+        "docvqa": 1
+      }
+    },
+    "scene_text_recognition": {
+      "n": 250,
+      "available": 6086,
+      "sources": {
+        "textvqa": 122,
+        "stvqa": 87,
+        "ocrvqa": 32,
+        "iconqa": 7,
+        "docvqa": 2
+      }
+    },
+    "document_qa": {
+      "n": 250,
+      "available": 1338,
+      "sources": {
+        "visualmrc": 123,
+        "docvqa": 66,
+        "tatqa": 50,
+        "infovqa": 7,
+        "scienceqa_image": 2,
+        "ocrvqa": 1,
+        "textvqa": 1
+      }
+    },
+    "document_text_reading": {
+      "n": 250,
+      "available": 4333,
+      "sources": {
+        "docvqa": 125,
+        "ocrvqa": 81,
+        "visualmrc": 28,
+        "textvqa": 9,
+        "infovqa": 4,
+        "stvqa": 2,
+        "tatqa": 1
+      }
+    },
+    "chart_reading": {
+      "n": 250,
+      "available": 15866,
+      "sources": {
+        "dvqa": 125,
+        "plotqa": 78,
+        "chartqa": 41,
+        "tabmwp": 3,
+        "docvqa": 2,
+        "infovqa": 1
+      }
+    },
+    "table_lookup": {
+      "n": 250,
+      "available": 1495,
+      "sources": {
+        "hitab": 73,
+        "docvqa": 71,
+        "tabmwp": 49,
+        "tatqa": 48,
+        "multihiertt": 6,
+        "infovqa": 2,
+        "scienceqa_image": 1
+      }
+    },
+    "key_information_extraction": {
+      "n": 168,
+      "available": 168,
+      "sources": {
+        "cord": 78,
+        "sroie": 76,
+        "funsd": 12,
+        "docvqa": 1,
+        "infovqa": 1
+      }
+    }
+  },
+  "bundle": {
+    "arm": "eval",
+    "rows": 2168,
+    "missing_images": 0,
+    "intervention": 2168,
+    "replay": 0,
+    "shards": 3,
+    "bytes": 501780247,
+    "max_side": 1792,
+    "by_capability": {
+      "chart_reasoning": 250,
+      "table_reasoning": 250,
+      "diagram_and_infographic_understanding": 250,
+      "scene_text_recognition": 250,
+      "document_qa": 250,
+      "document_text_reading": 250,
+      "chart_reading": 250,
+      "table_lookup": 250,
+      "key_information_extraction": 168
+    },
+    "by_source": {
+      "docvqa": 269,
+      "dvqa": 176,
+      "tabmwp": 168,
+      "plotqa": 158,
+      "visualmrc": 153,
+      "tatqa": 144,
+      "textvqa": 132,
+      "ai2d": 123,
+      "figureqa": 115,
+      "ocrvqa": 114,
+      "hitab": 97,
+      "infovqa": 95,
+      "stvqa": 89,
+      "cord": 78,
+      "sroie": 76,
+      "multihiertt": 68,
+      "scienceqa_image": 48,
+      "chartqa": 44,
+      "funsd": 12,
+      "iconqa": 9
+    }
+  }
+}