Improve dataset card: Update task category, add tags, and add sample usage

This PR updates the dataset card for "Forgotten Polygons: Multimodal Large Language Models are Shape-Blind" to enhance its accuracy and usability.

Key changes include:
- Updating the `task_categories` metadata from `image-classification` to `image-text-to-text`, which more accurately reflects the dataset's use in evaluating Multimodal Large Language Models (MLLMs) on visual-mathematical reasoning tasks.
- Adding relevant `tags` to improve discoverability, such as `mllm`, `multimodal`, `geometric-reasoning`, `shape-recognition`, `visual-question-answering`, `chain-of-thought`, `mathematics`, and `reasoning`.
- Including `language: - en` in the metadata.
- Introducing a "Sample Usage" section with a code snippet directly from the associated GitHub repository, demonstrating how to run an evaluation using the dataset.

These updates provide clearer context for users and improve the dataset's overall documentation on the Hugging Face Hub.

README.md

@@ -1,4 +1,17 @@
 ---
+task_categories:
+- image-text-to-text
+tags:
+- multimodal
+- mllm
+- geometric-reasoning
+- visual-question-answering
+- shape-recognition
+- chain-of-thought
+- mathematics
+- reasoning
+language:
+- en
 dataset_info:
   features:
   - name: shape
@@ -40,8 +53,6 @@ configs:
     path: data/heptagons_with_visual_cues-*
   - split: arrow_on_plus_with_visual_cues
     path: data/arrow_on_plus_with_visual_cues-*
-task_categories:
-- image-classification
 library_name:
 - pytorch
 ---
@@ -56,6 +67,19 @@ This dataset is part of the work **"Forgotten Polygons: Multimodal Large Languag
 
 This dataset is designed to evaluate the shape understanding capabilities of Multimodal Large Language Models (MLLMs).
 
+## Sample Usage
+
+This dataset is designed to be used with the evaluation code provided in the [GitHub Repository](https://github.com/rsinghlab/Shape-Blind/tree/main). To evaluate MLLMs on various tasks using this dataset, follow the instructions in the `evaluation` folder of the repository.
+
+For example, to run a shape identification task using LLaVA-1.5:
+
+```bash
+# Navigate to the 'evaluation' folder in the cloned GitHub repository
+cd Shape-Blind/evaluation
+# Run the evaluation script
+python3 evaluate_MLLMs.py --model_version llava-1.5 --task shape_id --dataset_size full
+```
+
 ## Dataset Splits
 
 Each split corresponds to a different reasoning task and shape identification challenge.