TAGCOS: Task-agnostic Gradient Clustered Coreset Selection for Instruction Tuning Data.
Jipeng ZhangYaxuan QinRenjie PiWeizhong ZhangRui PanTong ZhangPublished in: CoRR (2024)
Keyphrases
- data sets
- data collection
- data analysis
- image data
- synthetic data
- training data
- data mining techniques
- database
- small number
- statistical methods
- data processing
- data sources
- raw data
- neural network
- website
- probability distribution
- statistical analysis
- high dimensional data
- computer systems
- historical data
- original data
- attribute values
- genetic algorithm
- high quality
- input data
- bayesian networks
- prior knowledge
- data points