Jun 16, 2024 · We propose a single-step sampling calibration on activations and adapt group-wise quantization on weights for low-bit quantization.
scholar.google.com › citations
Jun 16, 2024 · In this research, we pioneer the investigation of quantizing a transformer-only diffusion model without any optimizations.
Jun 16, 2024 · By analysing challenges in quantizing activations and weights for diffusion transformers, we propose a single-step sampling calibration on ...
Jun 17, 2024 · This paper analyzes the process of quantizing Diffusion Transformers, which are a type of machine learning model used for tasks like image ...
Nov 13, 2024 · After analyzing the major challenges for quantizing diffusion transformers, we design an improved quantization scheme: "ViDiT-Q": Video and ...
This paper proposes PTQ4DiT, a novel Post-training Quantization (PTQ) method for Diffusion. Transformers (DiTs). Our analysis identifies the primary challenges ...
Jun 4, 2024 · After ana- lyzing the major challenges for quantizing diffusion transformers, we design an improved quantization scheme: ViDiT-Q (Video ...
People also ask
What is a diffusion transformer?
What is the difference between diffusion transformer and unet?
Sep 15, 2024 · To further enhance the performance of the quantized DiT, we adopt the layer-wise grid search strategy to optimize the smoothing factor.
... Quantization For Vision Transformer With Mixed Granularity Reconstruction, [Paper]; (arXiv 2024.06) An Analysis on Quantizing Diffusion Transformers, [Paper] ...
Nov 21, 2024 · New method called TaQ-DiT for quantizing diffusion transformers · Addresses accuracy loss in low-bit quantization of diffusion models · Uses time- ...