Recent MLLMs such as Qwen2-VL can match or beat CLIP-style models on several image classification benchmarks, with gains driven mainly by stronger LLMs and more diverse training data.
Objectnet: A large-scale bias-controlled dataset for pushing the limits of object recognition models.NeurIPS, 32, 2019
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CV 1years
2024 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
Revisiting MLLMs: An In-Depth Analysis of Image Classification Abilities
Recent MLLMs such as Qwen2-VL can match or beat CLIP-style models on several image classification benchmarks, with gains driven mainly by stronger LLMs and more diverse training data.