Split JSONL datasets into subsets by specified ratios.
npx skills add https://github.com/cas-bigdatalab/piflow --skill data-splitter
Please help me install this Agent Skill. Skill: data_splitter Source: https://github.com/cas-bigdatalab/piflow/tree/main/workspace/skills/data_splitter Command: npx skills add https://github.com/cas-bigdatalab/piflow --skill data-splitter
This Skill simplifies the process of splitting JSONL datasets into smaller subsets based on specified ratios, making it easier to manage and analyze data.
Use the data_splitter skill to split the 'data.jsonl' file into training, validation, and test sets with ratios 0.7, 0.2, and 0.1 respectively.
High-intent search queries and answers about installing and using this skill.
To split a JSONL dataset into training and test sets, this Skill divides the data into multiple subsets based on specified ratios, such as 0.7 for training and 0.3 for testing. It requires Python and specific command-line arguments for execution.
Stratified splitting for JSONL data is a mechanism that maintains the category distribution of a specific field when dividing the dataset. This ensures that the resulting training, validation, and test sets accurately reflect the original data's proportions.
Yes, you can create validation sets with random splitting from a JSONL file. This Skill supports random splitting to ensure better data distribution across your training, validation, and test subsets for machine learning workflows.
Yes, you need Python to manage dataset splitting for machine learning using this Skill. It is designed as a data preprocessing task executed through Python and specific command-line arguments to generate the required subsets.
qinyan-aiProcess Neuropixels datasets through preprocessing, drift correction, spike sorting, and curation.
henrycashe26Diagnose machine learning training runs by interpreting loss curves and validation metrics.
gehtalexeyFilter candidate records from enriched CSV by past candidates and blacklisted companies.
AMGrobelnikValidate JSON files against predefined schemas and generate full, mini, and preview outputs.
