Finalize dataset card

README.md

@@ -1,5 +1,5 @@
 ---
-license: other
+license: cc-by-nc-4.0
 task_categories:
   - visual-question-answering
   - image-text-to-text
@@ -10,7 +10,8 @@ tags:
   - multimodal
   - procedural-reasoning
   - benchmark
-pretty_name: Embedded-Device Tutorial Next-Step Benchmark
+  - computer-vision
+pretty_name: EmbedCopilot-Bench
 size_categories:
   - n<1K
 configs:
@@ -20,57 +21,48 @@ configs:
         path: data/test-*.parquet
 ---
 
-# Dataset Card for Embedded-Device Tutorial Next-Step Benchmark
+# EmbedCopilot-Bench
 
-## Dataset Summary
+EmbedCopilot-Bench is a multimodal benchmark for evaluating whether vision-language models can generate the next concrete instruction in embedded-device tutorials. Each example is built from consecutive tutorial action steps: the current action step provides the visual state and completed-step history, while the following action step provides the task query and ground-truth next-step instruction.
 
-This benchmark evaluates whether multimodal models can infer the next concrete instruction in embedded-device tutorials. Each sample is a consecutive action-step pair: the anchor action provides the current visual state, and the target action provides the instruction that the model must explain.
-
-The benchmark supports two settings:
-
-- `text_only`: completed-step history plus target action summary.
-- `with_images`: the same text context plus current device photo and/or computer screenshot.
+- Dataset repository: https://huggingface.co/datasets/X-EASys/EmbedCopilot-Bench
+- Code and evaluation scripts: https://github.com/X-EASys/EmbedCopilot-Bench
 
 ## Dataset Structure
 
-The `test` split contains 216 benchmark samples from 33 selected tutorial videos. The source videos contain 249 action steps; pairing consecutive action steps within each video yields 216 samples.
-
-Columns:
-
-- `question_id`: stable sample id.
-- `device_image`: anchor device/hardware context, rendered as an HF `Image` column when available.
-- `screenshot_image`: anchor computer screenshot context, rendered as an HF `Image` column when available.
-- `history_summaries`: completed-step history, exactly the original prediction input from `target_step["step_so_far"].strip()`.
-- `task_query`: target action summary, exactly the original prediction input from `target_step["summary"].strip()`.
-- `answer`: ground-truth immediate instruction text curated in the release staging file `data/samples.jsonl`.
-- `capability`: comma-joined labels such as `hw_assembly`, `env_sw_setup`, and `code_impl`.
-- `video_id`, `anchor_step_index`, `target_step_index`: minimal provenance fields.
-
-Prompt text is not duplicated as a data column. The accompanying GitHub code constructs the exact text-only and multimodal prompts from `history_summaries` and `task_query`.
+The released split contains 216 test examples. The split is stored as 12 Parquet shards under `data/test-*.parquet` so that the Hugging Face Dataset Viewer can load the benchmark reliably. Together, the shards form one `test` split.
 
-In the paper prompt notation, `History Summaries` maps to `history_summaries`, `Task Query` maps to `task_query`, and visual context maps to `device_image` plus `screenshot_image` from the anchor/current-state action step.
+Each row contains:
 
-## Usage
+| Column | Type | Description |
+| --- | --- | --- |
+| `question_id` | string | Unique sample identifier. |
+| `device_image` | image | Device-view image for the current tutorial step. |
+| `screenshot_image` | image or null | Optional software/screen-view image for the current tutorial step. |
+| `history_summaries` | string | Summary of tutorial steps completed before the target step. |
+| `task_query` | string | User-facing query for the target next step. |
+| `answer` | string | Ground-truth next-step instruction. |
+| `capability` | string | Coarse capability label, for example hardware assembly or software setup. |
+| `video_id` | string | Source tutorial video identifier. |
+| `anchor_step_index` | string | Current action step index within the source tutorial. |
+| `target_step_index` | string | Target next action step index within the source tutorial. |
 
-Dataset repository: https://huggingface.co/datasets/X-EASys/EmbedCopilot-Bench
+## Evaluation Protocol
 
-```python
-from datasets import load_dataset
+The benchmark prompt follows the paper's answer-generation setting: models receive the completed-step history, the visual context image or images, and the target task query, then generate the immediate method or instruction needed to accomplish the next step. The accompanying code repository provides scripts and a notebook for running model inference and judge-based evaluation with the released fields.
 
-ds = load_dataset("X-EASys/EmbedCopilot-Bench")
-print(ds["test"][0])
-```
+## Intended Use
 
-## Benchmark Protocol
+This dataset is intended for research on multimodal procedural reasoning, embedded-device assistance, tutorial understanding, and evaluation of vision-language models on hardware/software workflows. It is designed as an evaluation benchmark rather than a training corpus.
 
-For each video, sort steps with `is_action == "action"` by `step_index`, then pair consecutive action steps. The earlier action is the anchor/current state; the later action is the target. A model predicts the concrete next-step instruction. Evaluation compares the prediction to `answer` using the published rubric.
+## ESR Tasks
 
-For official GPT baselines, use the OpenAI Responses API directly with `OPENAI_API_KEY` only. Send `device_image` and `screenshot_image` as separate `input_image` payloads built from the original image bytes. Do not resize, crop, compress, collage, or convert images before API submission.
+The code repository also includes ESR task materials for reproducibility, including task-level README files and Wokwi project files where applicable. These files document the expected executable behavior used in the evaluation setting.
 
-## Known Issues
+## License
 
-The final 216-sample payload has no missing image references. License and redistribution terms for tutorial-derived frames must be finalized before public release.
+The dataset is released under the Creative Commons Attribution-NonCommercial 4.0 International license (CC BY-NC 4.0). The accompanying code repository may include code and scripts under the license specified in that repository.
 
-## Licensing
+## Citation
 
-License terms must be finalized before public release. Use `license: other` until media redistribution rights and final dataset terms are confirmed.
+If you use EmbedCopilot-Bench, please cite the corresponding EmbedCopilot paper.