|
- GitHub - hkust-nlp deita: Deita: Data-Efficient Instruction Tuning for . . .
Deita is an open-sourced project designed to facilitate Automatic Data Selection for instruction tuning in Large Language Models (LLMs) It includes: Open-sourced Toolkits for automatic data selection in instruction tuning; Deita Datasets: A series of extremely lightweight, high-quality alignment SFT data We release 6k-sized and 10k-sized
- Delta Air Lines - Airline Tickets and Airfare to Worldwide . . .
Delta Air Lines Book a trip Check in, change seats, track your bag, check flight status, and more
- Fine-tuning Large Language Models on a budget with DEITA - Mantis NLP
DEITA (Data-Efficient Instruction Tuning for Alignment) studies an automatic data selection process by first quantifying the data quality based on complexity, quality and diversity And second, selecting across the best potential combination from an open-source dataset that would fit into the budget you allocate to tune your own LLM
- DEITA - Distilabel Docs - Argilla
DEITA (Data-Efficient Instruction Tuning for Alignment) studies an automatic data selection process by first quantifying the data quality based on complexity, quality and diversity And second, selecting across the best potential combination from an open-source dataset that would fit into the budget you allocate to tune your own LLM
- hkust-nlp deita-6k-v0 · Datasets at Hugging Face
Deita is an open-sourced project designed to facilitate Automatic Data Selection for instruction tuning in Large Language Models (LLMs) This dataset includes 6k of lightweight, high-quality alignment SFT data, mainly automatically selected from the following datasets:
- arXiv:2312. 15685v2 [cs. CL] 16 Apr 2024
Published as a conference paper at ICLR 2024 WHAT MAKES GOOD DATA FOR ALIGNMENT? A COMPREHENSIVE STUDY OF AUTOMATIC DATA SELECTION IN INSTRUCTION TUNING Wei Liu ∗1 Weihao Zeng 2 Keqing He3 Yong Jiang4 Junxian He5 1ShanghaiTech University 2Beijing University of Posts and Telecommunications 3Meituan 4Alibaba Group 5The Hong Kong University of Science and Technology liuwei4@shanghaitech edu cn
- What Makes Good Data for Alignment? A Comprehensive Study of. . .
We present Deita (short for Data-Efficient Instruction Tuning for Alignment), a series of models fine-tuned from LLaMA models using data samples automatically selected with our proposed approach
- deita README. md at main · hkust-nlp deita - GitHub
Deita: Data-Efficient Instruction Tuning for Alignment [ICLR2024] - hkust-nlp deita
|
|
|