Upload folder using huggingface_hub

README.md

@@ -0,0 +1,92 @@
+---
+license: other
+license_name: mixed-original-licenses
+license_link: https://github.com/brendenlake/omniglot
+task_categories:
+  - image-classification
+  - visual-question-answering
+language:
+  - en
+tags:
+  - omniglot
+  - pacs
+  - vision-language
+  - visual-invariance
+  - geometric-reasoning
+  - domain-generalization
+  - arxiv:2604.01848
+  - serge
+pretty_name: SERGE
+size_categories:
+  - 1K<n<10K
+---
+
+# SERGE: SEmantic Richness & Geometric Evaluation
+
+SERGE is the evaluation dataset released with the paper [*Semantic Richness or Geometric Reasoning? The Fragility of VLM's Visual Invariance*](https://arxiv.org/abs/2604.01848). It contains the canonical (untransformed) images used to test whether VLMs judge object identity under rotation, scale, and identity transformations through geometric reasoning or through familiarity with semantically rich content.
+
+- **Paper:** [arXiv:2604.01848](https://arxiv.org/abs/2604.01848)
+- **Code:** [github.com/XThomasBU/semantic_richness](https://github.com/XThomasBU/semantic_richness)
+
+## Dataset overview
+
+| Dataset | # Samples | Description |
+| --- | --- | --- |
+| Omniglot | 1,623 | Handwritten characters across 50 scripts |
+| Times New Roman | 52 | Printed English alphabet (A–Z, a–z) |
+| Handwritten English | 52 | Handwritten English alphabet |
+| PACS | 800 | 4 domains × 200 images each |
+
+Together these span a semantic-richness spectrum used in the paper: symbolic sketches (Omniglot) → semantic sketches (Handwritten English, Times New Roman) → natural photos and art (PACS).
+
+### Directory layout
+
+```text
+omniglot/images_all/{Script}/characterXX/*.png   # 50 official scripts + hand_english
+times_new_roman/characterXX/image.png
+pacs/{domain}_{label}_{hf_index}.png
+pacs/selected_indices.json    # exact PACS indices sampled, for reproducibility
+manifest.json                 # per-script / per-domain counts
+```
+
+`hand_english` (the Handwritten English row above) is stored under `omniglot/images_all/hand_english` alongside the 50 official Omniglot scripts. `pacs` filenames encode metadata, e.g. `art_painting_0_101.png` → domain=`art_painting`, label=`0` (dog), source index=`101`.
+
+## Loading the dataset
+
+```python
+from huggingface_hub import snapshot_download
+from datasets import load_dataset
+
+local_dir = snapshot_download(repo_id="XThomasBU/SERGE", repo_type="dataset")
+
+omniglot = load_dataset("imagefolder", data_dir=f"{local_dir}/omniglot/images_all", split="train")
+times_new_roman = load_dataset("imagefolder", data_dir=f"{local_dir}/times_new_roman", split="train")
+pacs = load_dataset("imagefolder", data_dir=f"{local_dir}/pacs", split="train")
+```
+
+## Citation
+
+If you use SERGE, please cite the paper:
+
+```bibtex
+@misc{qiu2026semantic,
+  title={Semantic Richness or Geometric Reasoning? The Fragility of VLM's Visual Invariance},
+  author={Qiu, Jason and Meurer, Zachary and Thomas, Xavier and Ghadiyaram, Deepti},
+  year={2026},
+  eprint={2604.01848},
+  archivePrefix={arXiv},
+  primaryClass={cs.CV},
+  url={https://arxiv.org/abs/2604.01848}
+}
+```
+
+and the original source datasets:
+
+> Brenden M. Lake, Ruslan Salakhutdinov, and Joshua B. Tenenbaum. Human-level concept learning through probabilistic program induction. *Science*, 350(6266):1332–1338, 2015. doi: 10.1126/science.aab3050. URL https://www.science.org/doi/abs/10.1126/science.aab3050.
+>
+> Sujay Mann. Handwritten english characters and digits, 2024. URL https://www.kaggle.com/datasets/sujaymann/handwritten-english-characters-and-digits. Accessed: 2026-03-29.
+>
+> Samuel Yu, Peter Wu, Paul Pu Liang, Ruslan Salakhutdinov, and Louis-Philippe Morency. Pacs: A dataset for physical audiovisual commonsense reasoning, 2022. URL https://arxiv.org/abs/2203.11130.
+>
+> Stanley Morison and Victor Lardent. Times new roman. https://en.wikipedia.org/wiki/Times_New_Roman, 1932. Typeface originally commissioned for The Times newspaper. Accessed: 2026-03-30.
+

manifest.json

@@ -0,0 +1,241 @@
+{
+  "dataset_name": "SERGE",
+  "full_name": "SEmantic Richness & Geometric Evaluation",
+  "description": "SERGE evaluation image subset used by VLM_BLIND experiments for scale, identity/spatial, and rotation invariance tests.",
+  "omniglot": {
+    "scripts": 51,
+    "characters": 1675,
+    "images": 1675,
+    "strokes_per_char": 1,
+    "script_counts": {
+      "Alphabet_of_the_Magi": {
+        "characters": 20,
+        "images": 20
+      },
+      "Angelic": {
+        "characters": 20,
+        "images": 20
+      },
+      "Anglo-Saxon_Futhorc": {
+        "characters": 29,
+        "images": 29
+      },
+      "Arcadian": {
+        "characters": 26,
+        "images": 26
+      },
+      "Armenian": {
+        "characters": 41,
+        "images": 41
+      },
+      "Asomtavruli_(Georgian)": {
+        "characters": 40,
+        "images": 40
+      },
+      "Atemayar_Qelisayer": {
+        "characters": 26,
+        "images": 26
+      },
+      "Atlantean": {
+        "characters": 26,
+        "images": 26
+      },
+      "Aurek-Besh": {
+        "characters": 26,
+        "images": 26
+      },
+      "Avesta": {
+        "characters": 26,
+        "images": 26
+      },
+      "Balinese": {
+        "characters": 24,
+        "images": 24
+      },
+      "Bengali": {
+        "characters": 46,
+        "images": 46
+      },
+      "Blackfoot_(Canadian_Aboriginal_Syllabics)": {
+        "characters": 14,
+        "images": 14
+      },
+      "Braille": {
+        "characters": 26,
+        "images": 26
+      },
+      "Burmese_(Myanmar)": {
+        "characters": 34,
+        "images": 34
+      },
+      "Cyrillic": {
+        "characters": 33,
+        "images": 33
+      },
+      "Early_Aramaic": {
+        "characters": 22,
+        "images": 22
+      },
+      "Futurama": {
+        "characters": 26,
+        "images": 26
+      },
+      "Ge_ez": {
+        "characters": 26,
+        "images": 26
+      },
+      "Glagolitic": {
+        "characters": 45,
+        "images": 45
+      },
+      "Grantha": {
+        "characters": 43,
+        "images": 43
+      },
+      "Greek": {
+        "characters": 24,
+        "images": 24
+      },
+      "Gujarati": {
+        "characters": 48,
+        "images": 48
+      },
+      "Gurmukhi": {
+        "characters": 45,
+        "images": 45
+      },
+      "Hebrew": {
+        "characters": 22,
+        "images": 22
+      },
+      "Inuktitut_(Canadian_Aboriginal_Syllabics)": {
+        "characters": 16,
+        "images": 16
+      },
+      "Japanese_(hiragana)": {
+        "characters": 52,
+        "images": 52
+      },
+      "Japanese_(katakana)": {
+        "characters": 47,
+        "images": 47
+      },
+      "Kannada": {
+        "characters": 41,
+        "images": 41
+      },
+      "Keble": {
+        "characters": 26,
+        "images": 26
+      },
+      "Korean": {
+        "characters": 40,
+        "images": 40
+      },
+      "Latin": {
+        "characters": 26,
+        "images": 26
+      },
+      "Malay_(Jawi_-_Arabic)": {
+        "characters": 40,
+        "images": 40
+      },
+      "Malayalam": {
+        "characters": 47,
+        "images": 47
+      },
+      "Manipuri": {
+        "characters": 40,
+        "images": 40
+      },
+      "Mkhedruli_(Georgian)": {
+        "characters": 41,
+        "images": 41
+      },
+      "Mongolian": {
+        "characters": 30,
+        "images": 30
+      },
+      "N_Ko": {
+        "characters": 33,
+        "images": 33
+      },
+      "Ojibwe_(Canadian_Aboriginal_Syllabics)": {
+        "characters": 14,
+        "images": 14
+      },
+      "Old_Church_Slavonic_(Cyrillic)": {
+        "characters": 45,
+        "images": 45
+      },
+      "Oriya": {
+        "characters": 46,
+        "images": 46
+      },
+      "Sanskrit": {
+        "characters": 42,
+        "images": 42
+      },
+      "Sylheti": {
+        "characters": 28,
+        "images": 28
+      },
+      "Syriac_(Estrangelo)": {
+        "characters": 23,
+        "images": 23
+      },
+      "Syriac_(Serto)": {
+        "characters": 23,
+        "images": 23
+      },
+      "Tagalog": {
+        "characters": 17,
+        "images": 17
+      },
+      "Tengwar": {
+        "characters": 25,
+        "images": 25
+      },
+      "Tibetan": {
+        "characters": 42,
+        "images": 42
+      },
+      "Tifinagh": {
+        "characters": 55,
+        "images": 55
+      },
+      "ULOG": {
+        "characters": 26,
+        "images": 26
+      },
+      "hand_english": {
+        "characters": 52,
+        "images": 52
+      }
+    },
+    "notes": "Official Omniglot background+evaluation alphabets (first stroke only), plus local hand_english if present."
+  },
+  "times_new_roman": {
+    "characters": 52,
+    "images": 52,
+    "source": "DATA/times_new_roman"
+  },
+  "pacs": {
+    "source": "flwrlabs/pacs",
+    "split": "train",
+    "samples_per_domain": 200,
+    "seed": 42,
+    "total_images": 800,
+    "domains": {
+      "art_painting": 200,
+      "cartoon": 200,
+      "photo": 200,
+      "sketch": 200
+    }
+  },
+  "layout": {
+    "omniglot/images_all/{Script}/characterXX/*.png": "first N strokes per character",
+    "times_new_roman/characterXX/image.png": "Latin control",
+    "pacs/{domain}_{label}_{hf_index}.png": "flat PACS eval subset (rotation + scale)"
+  }
+}