Diffusion models tolerate pruning up to 90% of training data without FID degradation, and cluster-center selection in CLIP/DINO embedding space beats established gradient-based pruning methods.
Sparsegpt: Mas- sive language models can be accurately pruned in one-shot
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.LG 1years
2024 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
Data Pruning in Generative Diffusion Models
Diffusion models tolerate pruning up to 90% of training data without FID degradation, and cluster-center selection in CLIP/DINO embedding space beats established gradient-based pruning methods.