423A35C7/LLaMA-Factory
1
0
Fork 0
mirror of https://github.com/hiyouga/LLaMA-Factory.git synced 2025-12-15 11:20:35 +08:00
Code Issues Packages Projects Releases Wiki Activity

fix #2735

Browse Source
This commit is contained in:
hiyouga
2024-03-07 16:15:53 +08:00
parent 2185855bdb
commit f74f804a71
4 changed files with 41 additions and 30 deletions
Download Patch File Download Diff File Expand all files Collapse all files

6
src/llmtuner/model/patcher.py
View File

@@ -286,9 +286,9 @@ def patch_config(
init_kwargs["torch_dtype"] = model_args.compute_dtype
if not is_deepspeed_zero3_enabled():
init_kwargs["low_cpu_mem_usage"] = True
if "device_map" not in init_kwargs:
init_kwargs["device_map"] = {"": get_current_device()} if is_trainable else "auto"
init_kwargs["low_cpu_mem_usage"] = model_args.low_cpu_mem_usage
if "device_map" not in init_kwargs: # quant models cannot use auto device map
init_kwargs["device_map"] = model_args.device_map or {"": get_current_device()}
def patch_model(