Method Drift›Parameter-efficient fine-tuning (LoRA family)
CLIP-Adapter
CLIP-Adapter: Better Vision-Language Models with Feature AdaptersParameter-efficient fine-tuning (LoRA family) · first seen Oct 9, 2021
superseded — cited as a baseline and beaten by newer methods
0 papers critique it · 3 beat it on benchmarks
Beaten on benchmarks
Head-to-head results where a newer method reports beating CLIP-Adapter. Values are copied from the source paper's tables — verify against the cited paper.
- Multi-Modal Parameter-Efficient Fine-tuning via Graph Neural Network
99.02 beats CLIP-Adapter · accuracy [Flowers102]
99.02 vs 93.90
- Multi-Modal Parameter-Efficient Fine-tuning via Graph Neural Network
92.63 beats CLIP-Adapter · accuracy [Oxford Pets]
92.63 vs 87.84
- Multi-Modal Parameter-Efficient Fine-tuning via Graph Neural Network
78.46 beats CLIP-Adapter · accuracy [Food101]
78.46 vs 78.25
- Quantified Task Misalignment to Inform PEFT: An Exploration of Domain Generalization and Catastrophic Forgetting in CLIP
A-CLIP beats CLIP-Adapter · ID Model Performance [Cars]
79.12 vs 70.60
- Quantified Task Misalignment to Inform PEFT: An Exploration of Domain Generalization and Catastrophic Forgetting in CLIP
A-CLIP beats CLIP-Adapter · ID Model Performance [FMoW-ID]
58.67 vs 44.43
- Quantified Task Misalignment to Inform PEFT: An Exploration of Domain Generalization and Catastrophic Forgetting in CLIP
A-CLIP beats CLIP-Adapter · ID Model Performance [16-Shot ImageNet]
66.96 vs 64.15
- ACE-LoRA: Graph-Attentive Context Enhancement for Parameter-Efficient Adaptation of Medical Vision-Language Models
ACE-LoRA beats CLIP-Adapter · Accuracy [Zero-shot classification (Radiology)]
49.80 vs 42.70
Newer alternatives
Recent methods in the same sub-problem, not yet superseded in the knowledge base.