Update OCR captcha dataset metadata for 9000-label snapshot

README.md

@@ -11,21 +11,21 @@ size_categories:
 
 # eKYB Captcha Labeled Dataset
 
-Generated at: 2026-02-26T13:02:56.462974+00:00
+Generated at: 2026-05-10T17:09:52Z
 
 ## Summary
 
-- Source labels CSV: `/Users/huynhthanhdat/Workspace/iNexus/eKYB/crawler/out/labeling_batches/all-captcha-labeling-dataset-20260226/labels.csv`
-- Rows total in CSV: `4300`
-- Rows with label: `500`
-- Rows unlabeled: `3800`
+- Source metadata CSV: `/Users/huynhthanhdat/Workspace/iNexus/eKYB/ocr-captcha-finetuned/datasets/hf_dataset/source/metadata.csv`
+- Rows total in source: `10954`
+- Rows with label: `9000`
+- Rows unlabeled: `1954`
 - Rows skipped (missing image): `0`
-- Exported labeled samples: `500`
+- Exported labeled samples: `9000`
 
 ## Splits
 
-- train: `450`
-- validation: `50`
+- train: `8100`
+- validation: `900`
 - test: `0`
 
 ## Files
@@ -33,27 +33,22 @@ Generated at: 2026-02-26T13:02:56.462974+00:00
 - HF `imagefolder` standard layout:
   - `train/*.png`
   - `train/metadata.csv`
-  - `validation/*.png` + `validation/metadata.csv` (if split > 0)
-  - `test/*.png` + `test/metadata.csv` (if split > 0)
+  - `validation/*.png`
+  - `validation/metadata.csv`
 
-Each split `metadata.csv` has columns:
+Each split `metadata.csv` keeps the existing remote columns:
 
 - `file_name` (file inside the split folder)
-- `text` (label from CSV, preserved as entered)
+- `text` (label, preserved as entered)
 - `id`
 - `source` (`canhantmdt` / `tracuunnt` / `unknown`)
-- `notes`
-- `image_file`
 - `source_path`
+- `notes`
 
 ## Load Example
 
 ```python
 from datasets import load_dataset
 
-# local
 ds = load_dataset("imagefolder", data_dir=".")
-
-# from hub
-# ds = load_dataset("imagefolder", data_dir="https://huggingface.co/datasets/dathuynh1108/captcha-to-text/resolve/main")
 ```

dataset_metadata.json

@@ -1,15 +1,15 @@
 {
-  "generated_at": "2026-02-26T13:02:56.462571+00:00",
-  "labels_csv": "/Users/huynhthanhdat/Workspace/iNexus/eKYB/crawler/out/labeling_batches/all-captcha-labeling-dataset-20260226/labels.csv",
-  "rows_total": 4300,
-  "rows_labeled": 500,
-  "rows_unlabeled": 3800,
+  "generated_at": "2026-05-10T17:09:52Z",
+  "labels_csv": "/Users/huynhthanhdat/Workspace/iNexus/eKYB/ocr-captcha-finetuned/datasets/hf_dataset/source/metadata.csv",
+  "rows_total": 10954,
+  "rows_labeled": 9000,
+  "rows_unlabeled": 1954,
   "rows_missing_image": 0,
-  "exported_samples": 500,
+  "exported_samples": 9000,
   "split_stats": {
-    "train": 450,
-    "validation": 50,
+    "train": 8100,
+    "validation": 900,
     "test": 0
   },
   "repo_id": "dathuynh1108/captcha-to-text"
-}
+}