Sammenlign metoder
Gjennomgå de valgte metodene side om side; rader som avviker, er uthevet.
| Multi-hode selvoppmerksomhet× | LoRA og PEFT× | |
|---|---|---|
| Fagfelt | Dyp læring | Dyp læring |
| Familie | Machine learning | Machine learning |
| Opprinnelsesår≠ | 2017 | 2022 |
| Opphavsperson≠ | Vaswani, A. et al. | Hu, E. J. et al.; Lester, B. et al. |
| Type≠ | Attention mechanism (Transformer core) | Parameter-efficient fine-tuning of large pretrained models |
| Opprinnelig kilde≠ | Vaswani, A. et al. (2017). Attention Is All You Need. NeurIPS. link ↗ | Hu, E. J. et al. (2022). LoRA: Low-Rank Adaptation of Large Language Models. ICLR. link ↗ |
| Alias≠ | Öz-Dikkat ve Çok Başlı Dikkat (Multi-Head Self-Attention), öz-dikkat, multi-head attention, scaled dot-product attention | LoRA ve PEFT — Parametre Verimli İnce Ayar, Low-Rank Adaptation, parameter-efficient fine-tuning, prefix tuning |
| Relaterte | 5 | 5 |
| Sammendrag≠ | 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. | 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. |
| ScholarGateDatasett ↗ |
|
|