From 3d561b60ac90b3124466e83d82dac9192792fa6d Mon Sep 17 00:00:00 2001 From: binbin Deng <108676127+plusbang@users.noreply.github.com> Date: Mon, 15 Apr 2024 12:18:25 +0800 Subject: [PATCH] LLM: add `enable_xetla` parameter for `optimize_model` API (#10753) --- python/llm/src/ipex_llm/optimize.py | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-) diff --git a/python/llm/src/ipex_llm/optimize.py b/python/llm/src/ipex_llm/optimize.py index dc199c00..bef3dfca 100644 --- a/python/llm/src/ipex_llm/optimize.py +++ b/python/llm/src/ipex_llm/optimize.py @@ -253,7 +253,8 @@ def optimize_model(model, low_bit='sym_int4', optimize_llm=True, modules_to_not_ optimize_model=optimize_llm, modules_to_not_convert=modules_to_not_convert, cpu_embedding=cpu_embedding, - lightweight_bmm=lightweight_bmm) + lightweight_bmm=lightweight_bmm, + enable_xetla=kwargs.pop("enable_xetla", False)) # add save_low_bit to pretrained model dynamically import types model._bigdl_config = dict()