From 4cf1ccc43a69dc5ee9269261dc63295758fe8af1 Mon Sep 17 00:00:00 2001 From: Rahul Nair Date: Wed, 30 Oct 2024 18:56:46 -0800 Subject: [PATCH] Update DPO EADME.md (#12162) bitsanbytes multi backend is now available and is required , otherwise would error out saying that no cuda is available --- python/llm/example/GPU/LLM-Finetuning/DPO/README.md | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-) diff --git a/python/llm/example/GPU/LLM-Finetuning/DPO/README.md b/python/llm/example/GPU/LLM-Finetuning/DPO/README.md index 4bd41efdcc5..9fdf34cbbb0 100644 --- a/python/llm/example/GPU/LLM-Finetuning/DPO/README.md +++ b/python/llm/example/GPU/LLM-Finetuning/DPO/README.md @@ -19,7 +19,8 @@ conda activate llm pip install --pre --upgrade ipex-llm[xpu] --extra-index-url https://pytorch-extension.intel.com/release-whl/stable/xpu/us/ pip install transformers==4.36.0 datasets pip install trl peft==0.10.0 -pip install bitsandbytes +# Note, if you don't want to reinstall BNBs dependencies, append the `--no-deps` flag! +pip install --no-deps --force-reinstall 'https://github.com/bitsandbytes-foundation/bitsandbytes/releases/download/continuous-release_multi-backend-refactor/bitsandbytes-0.44.1.dev0-py3-none-manylinux_2_24_x86_64.whl' ``` ### 2. Configures OneAPI environment variables