diffalign.models.epsnet.diffusion.merge_graphs_in_batch¶

diffalign.models.epsnet.diffusion.merge_graphs_in_batch(batch1: Batch, batch2: Batch, device: device | None = None) → Batch[source]¶: Merge as [Q1,R1,Q2,R2,…] and attach graph_idx (even=query, odd=ref) and pair-level batch. All Data objects are moved to device beforehand to avoid CPU/CUDA mixing.