diffusers

716f2460 - Fix UniPC scheduler device mismatch when using offloading (#13489)

Fix UniPC scheduler device mismatch when using offloading (#13489) When model/CPU offloading is enabled, self.sigmas may reside on CPU while the sample tensor is on GPU. The multistep_uni_p_bh_update and multistep_uni_c_bh_update methods index self.sigmas without moving the result to the sample device, causing torch.stack(rks) to fail with "Expected all tensors to be on the same device". Move sigma values to the sample device immediately after indexing, ensuring all derived tensors (lambda, h, rk) stay on the correct device throughout the computation. Fixes #13488 Co-authored-by: Dhruv Nair <dhruv.nair@gmail.com>