مقایسهٔ روش‌ها

	توجه به خود چند-سری	تنظیم دقیق BERT	لورا و PEFT
حوزه	یادگیری عمیق	یادگیری عمیق	یادگیری عمیق
خانواده	Machine learning	Machine learning	Machine learning
سال پیدایش	2017	2019	2022
پدیدآور	Vaswani, A. et al.	Devlin, J. et al.	Hu, E. J. et al.; Lester, B. et al.
نوع	Attention mechanism (Transformer core)	Transfer learning (fine-tuning a pre-trained transformer)	Parameter-efficient fine-tuning of large pretrained models
منبع بنیادین	Vaswani, A. et al. (2017). Attention Is All You Need. NeurIPS.	Devlin, J., Chang, M.-W., Lee, K. & Toutanova, K. (2019). BERT: Pre-training of Deep Bidirectional Transformers for Language Understanding. NAACL.	Hu, E. J. et al. (2022). LoRA: Low-Rank Adaptation of Large Language Models. ICLR.
نام‌های دیگر	Öz-Dikkat ve Çok Başlı Dikkat (Multi-Head Self-Attention), öz-dikkat, multi-head attention, scaled dot-product attention	BERT İnce Ayar (Fine-Tuning), BERT ince ayar, fine-tuning BERT, transfer learning with BERT	LoRA ve PEFT — Parametre Verimli İnce Ayar, Low-Rank Adaptation, parameter-efficient fine-tuning, prefix tuning
مرتبط	5	5	5
خلاصه	Multi-head self-attention, introduced by Vaswani and colleagues in 2017, is the mechanism that lets every position in a sequence compute its relationship to all other positions in parallel. It is the core of the Transformer architecture and the foundation underneath BERT, GPT, and T5.	BERT fine-tuning, building on the BERT model introduced by Devlin and colleagues in 2019, re-trains a pre-trained BERT model on a small labelled dataset for a target task such as classification, named-entity recognition, or question answering. Through transfer learning it reaches high performance even with relatively little task-specific data.	LoRA (Low-Rank Adaptation), introduced by Hu et al. in 2022, and the broader family of parameter-efficient fine-tuning (PEFT) methods adapt large pretrained language models to new tasks by training only a small number of extra parameters instead of every weight in the model. This makes fine-tuning possible with far less GPU memory and compute while leaving the original model largely untouched.