Upload README.md with huggingface_hub

README.md

@@ -1,23 +1,33 @@
 ---
-dataset_info:
-  features:
-  - name: source_index
-    dtype: int64
-  - name: image
-    dtype: image
-  - name: user_content
-    dtype: string
-  - name: assistant_content
-    dtype: string
-  splits:
-  - name: train
-    num_bytes: 500071258
-    num_examples: 10000
-  download_size: 487440908
-  dataset_size: 500071258
-configs:
-- config_name: default
-  data_files:
-  - split: train
-    path: data/train-*
+task_categories:
+- image-text-to-text
+tags:
+- chart-to-code
+- code-generation
+- multimodal
+pretty_name: VinciCoder Chart2Code 0-10k (seed 0)
 ---
+
+# VinciCoder Chart2Code, rows 0-10000 of seed 0
+
+10,000 chart-to-code examples (chart image, instruction, reference matplotlib program) drawn from
+the `chart2code_1.parquet` file of [DocTron-Hub/VinciCoder-1.6M-SFT](https://huggingface.co/datasets/DocTron-Hub/VinciCoder-1.6M-SFT)
+(revision `974a038fb6f695e7a935112d4df430926fc468b3`). All rights and terms are those of the source dataset and the works it collects.
+
+## Selection
+
+- The file's 163,186 rows are split by SHA-256 of the image: 1% held out (1,678 rows, never included here),
+  161,508 for training.
+- The training rows are shuffled once with `random.Random(0).shuffle(train_indices)`; this dataset is
+  positions `[0, 10000)` of that order, in that order. `permutation.json` holds the full order
+  (sha256 `bc4a30fffc10377e4c56b71dfbe5679a5e93e5d3723a392c4d7abcc3cc90aa34`), so `[10000, 10000 + N)` gives the next, non-overlapping chunk.
+- `manifest.json` records the source split and the sha256 of these indices (`f399c46e442a9895af38ed8c7e74a3d884f5059980c2a486f9a4d8ec3c32c8ff`).
+
+## Columns
+
+| Column | Meaning |
+|---|---|
+| `source_index` | Row number in the source `chart2code_1.parquet` |
+| `image` | Reference chart |
+| `user_content` | Instruction |
+| `assistant_content` | Reference answer (matplotlib code) |