This project develops an advanced AI model specialized in providing guidance on GDPR (General Data Protection Regulation) compliance.
By fine-tuning Google's Gemma 2B model using Direct Preference Optimization (DPO) and a GDPR-specific dataset,
we've created a powerful tool to assist organizations with data protection queries and regulatory compliance.
- Specialized in GDPR compliance and data protection regulations
- Utilizes DPO for precise alignment with GDPR principles
- Implements QLoRA for efficient and resource-friendly training
- Designed to provide accurate and relevant responses to GDPR-related inquiries
- Base Model: Google Gemma 2B
- Fine-tuning Method: Direct Preference Optimization (DPO)
- Training Dataset: sims2k/GDPR_QA_instruct_dataset
- Quantization: 4-bit quantization using QLoRA
-
QLoRa Fine-tuning
-
Code Reference
-
Related LLM project
-
Concept
- sims2k/GDPR_QA_instruct_dataset
- sims2k/GDPR_QA_instruct_eval_dataset
- (github)gdpr-dataset
- Is Your Policy Compliant?: A Deep Learning-based Empirical Study of Privacy Policies' Compliance with GDPR
- (구글코라이 블로그)구글의 최첨단 오픈 모델 '젬마(Gemma)'를 공개합니다
- Gemma 2 model card
- Encoder Only 와 Decoder Only 언어모델에 대한 고찰
- (youtube)Fine-tuning LLMs | w/ Example Code
- Low-Rank Adapter (LoRA) Explained
- (medium)Getting Started with Google's Gemma LLM using HuggingFace Libaries
- (DEVOCEAN) Gemma 한국어 요약 모델 파인튜닝 빠르게 해보기
- (DEVOCEAN)오픈소스 LLM에 새로운 표준을 제시할 구글 Gemma
- (blog)Gemma: Open Models Based on GeminiResearch and Technology 논문 리뷰
- Sherlock Holmes Q&A with Gemma fine tuning