docs(prepare_model_for_kbit_training): note the ~0.5–1 GB CUDA reserved overhead (#3265)#3267

Anai-Guo wants to merge 1 commit into

huggingface:mainhuggingface/peft:mainfrom

Anai-Guo:docs/prepare-for-kbit-memory-noteAnai-Guo/peft:docs/prepare-for-kbit-memory-noteCopy head branch name to clipboard