QLIP predicts the quality of the image a text prompt will generate and then assigns low, medium, or high bit precision per layer and denoising step, reducing computational cost while matching or improving FID.
Improving image generation with better captions
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
citation-role summary
background 1
citation-polarity summary
fields
cs.CV 1years
2025 1verdicts
CONDITIONAL 1roles
background 1polarities
unclear 1representative citing papers
citing papers explorer
-
Text Embedding Knows How to Quantize Text-Guided Diffusion Models
QLIP predicts the quality of the image a text prompt will generate and then assigns low, medium, or high bit precision per layer and denoising step, reducing computational cost while matching or improving FID.