ConceptDiT / Diffusion Transformer
Diffusion Transformer
A design that introduces the Transformer architecture into diffusion or flow-matching generative backbones, replacing the traditional U-Net convolutional backbone with Transformer blocks, modeling long-range dependencies through self-attention, and injecting timestep and conditioning information via adaptive normalization.
Timeline (last 90 days)
Nov 3
FlowBatt is a battery degradation prediction framework based on conditional generative models, using a flow matching method combined with a Diffusion Transformer (DiT) backbone to generate future SOH trajectories conditioned on early-cycle capacity data.
Unverified50%