423A35C7/LLaMA-Factory
1
0
Fork 0
mirror of https://github.com/hiyouga/LLaMA-Factory.git synced 2025-08-04 12:42:51 +08:00
Code Issues Packages Projects Releases Wiki Activity

Update patcher.py

Browse Source 
Former-commit-id: a950f3b81de701f5f23ce3efa60ff0382bb40dfe
This commit is contained in:
hoshi-hiyouga 2024-04-16 17:29:19 +08:00 committed by GitHub
parent ce56ff22af
commit 48fb0be1b9
1 changed files with 3 additions and 2 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

5
src/llmtuner/model/patcher.py
View File

@ -133,7 +133,9 @@ def _configure_quantization(
if is_deepspeed_zero3_enabled():
raise ValueError("DeepSpeed ZeRO-3 is incompatible with quantized models.")
init_kwargs["device_map"] = {"": get_current_device()}
if model_args.quantization_device_map != "auto":
init_kwargs["device_map"] = {"": get_current_device()}
quantization_config: Dict[str, Any] = getattr(config, "quantization_config", None)
quant_method = quantization_config.get("quant_method", "")
@ -268,7 +270,6 @@ def _prepare_model_for_training(
# According to: https://github.com/huggingface/transformers/issues/28339
model.gradient_checkpointing_enable = MethodType(gradient_checkpointing_enable, model)
model.gradient_checkpointing_enable(gradient_checkpointing_kwargs={"use_reentrant": True})
# model.enable_input_require_grads()
setattr(model.config, "use_cache", False) # turn off when gradient checkpointing is enabled
logger.info("Gradient checkpointing enabled.")