Datasets:
| language: | |
| - da | |
| pretty_name: Handwritten Machine Text 2 | |
| size_categories: | |
| - 10K<n<100K | |
| task_categories: | |
| - image-classification | |
| # DanskScanning/handwritten-machine-2 | |
| Parquet dataset with 11,379 rows. Each row references an image file on the training storage volume. | |
| ## Labels | |
| - `0`: printed / machine-written (9,091 rows) | |
| - `1`: handwritten (2,288 rows) | |
| ## Parquet schema | |
| - `id`: string | |
| - `category`: int8 | |
| - `linux_path`: string | |
| The image bytes are not embedded. `linux_path` values use `/mnt/i_drev/Users/Filip/AI-training/training_data_25-09/labelled/` and require that volume to be mounted on the training server, matching the previous dataset format. | |