Fix sigma_last with use_flow_sigmas

[hlky]

What does this PR do?

sigma_last is concatenated in different places across schedulers so this was missed for some schedulers with the recent use_flow_sigmas, also missed with use_exponential_sigmas and use_beta_sigmas in DPMSolverMultistepInverseScheduler.

Fixes #10266

Who can review?

Anyone in the community is free to review the PR once the tests have passed. Feel free to tag
members/contributors who may be interested in your PR.

@yiyixuxu @vladmandic

[HuggingFaceDocBuilderDev]

The docs for this PR live here. All of your documentation changes will be reflected on that endpoint. The docs are available until 30 days after the last update.

[hlky]

After this is merged I'll add a test that checks the length of sigmas and timesteps for each scheduler and each use_*_sigmas option so we can catch this issue in the future and I'll standardise the code for concatenating sigma_last.

[yiyixuxu]

thanks!