starlight-apk/Qwen
1
0
Fork 0
mirror of https://github.com/QwenLM/Qwen.git synced 2026-05-21 00:45:48 +08:00
Code Issues Packages Projects Releases Wiki Activity

Merge pull request #964 from QwenLM/fix-finetune

Browse Source 
Fix bug of low_cpu_mem_usage in finetune.py.
This commit is contained in:
yang fan
2024-01-15 12:50:47 +08:00
committed by GitHub
parent ab109ced9f 73b34d4a9d
commit 204c2c59f4
1 changed files with 3 additions and 3 deletions
Download Patch File Download Diff File Expand all files Collapse all files

6
finetune.py
View File

@@ -291,9 +291,9 @@ def train():
): ):
raise RuntimeError("ZeRO3 is incompatible with LoRA when finetuning on base model.") raise RuntimeError("ZeRO3 is incompatible with LoRA when finetuning on base model.")
model_load_kwargs = {} model_load_kwargs = {
if deepspeed.is_deepspeed_zero3_enabled(): 'low_cpu_mem_usage': not deepspeed.is_deepspeed_zero3_enabled(),
model_load_kwargs['low_cpu_mem_usage'] = False }
# Set RoPE scaling factor # Set RoPE scaling factor
config = transformers.AutoConfig.from_pretrained( config = transformers.AutoConfig.from_pretrained(