[Bugfix] Fix getting vision features in Transformer Multimodal backend#32933

DarkLight1337 merged 2 commits intovllm-project:mainvllm-project/vllm:mainfrom

zucchini-nlp:get-image-features-v5-updatezucchini-nlp/vllm:get-image-features-v5-updateCopy head branch name to clipboard