fp_pt

# Dataset Card for fp_pt_split_hub ## Dataset Summary This dataset contains text records converted from local JSONL files. Each source record had a text field, which is exposed as the dataset's `text` column. Conversion cleanup: - HTML character references such as `'`, `&`, `<`, and `"` are decoded. - Conservative text cleanup removes transcript stage tags, common YouTube boilerplate, control/replacement characters, PDF percent-space artifacts, and raw LaTeX `tabular` wrappers. - Punctuation spacing is normalized for common extraction artifacts such as `A.Connella`, `ISSN:2153`, and `text.Next`. ## How to Load ```python from datasets import load_dataset ds = load_dataset("fp_pt_split_hub") ``` For a local checkout, use: ```python from datasets import load_dataset ds = load_dataset("/path/to/fp_pt_split_hub") ``` ## Dataset Stats | Split | Rows | | --- | ---: | | `train` | 14,522 | | `validation` | 147 | | `test` | 75 | ## Data Schema Columns: - `text` ## Licensing Information License: other