Tag: parameter-efficient fine-tuning

Researchers fine-tuned a Llama model with LoRA to extract agronomic triplets and build a GraphRAG-powered knowledge graph that outperforms standard ...

A new federated framework fine-tunes the Segment Anything Model for multi-organ segmentation by training only tiny low-rank adapters on quantized ...

Researchers have developed ValuePEFT, a parameter-efficient fine-tuning method that steers large language models toward explicitly specified human values with measurable ...

Researchers have introduced HGA-Net, a hypergraph-based adapter that connects related examples within each mini-batch and achieves near-perfect fine-grained recognition with ...

Researchers have developed Fed-A²LoRA, a federated fine-tuning method that halves trainable adapter parameters and cuts communication costs by up to ...

A matched-budget study of LoRA fine-tuning finds that genetic algorithms and other search optimizers offer no consistent advantage over well-tuned ...

Researchers have introduced PLS-ZO, a zeroth-order optimization framework that confines gradient-free perturbations to a low-dimensional latent subspace, dramatically reducing the ...

Researchers at Tarbiat Modares University have developed DuoDiT, a dual-stream architecture that adapts large diffusion transformers for image generation by ...

A new systematic study shows that low-rank adaptation can match or approach full fine-tuning of transformer models across four sentiment ...

Penn State researchers have developed a continual learning method that adaptively adjusts LoRA ranks layer by layer, helping large language ...