423A35C7/LLaMA-Factory
1
0
Fork 0
mirror of https://github.com/hiyouga/LLaMA-Factory.git synced 2025-08-05 21:22:50 +08:00
Code Issues Packages Projects Releases Wiki Activity

Update parser.py

Browse Source 
Former-commit-id: c00f0771a5ab2a422e0300dcb6d88e8609a3b997
This commit is contained in:
hiyouga 2024-04-16 18:09:31 +08:00
parent 8beb7a9239
commit 038d524fcd
1 changed files with 3 additions and 3 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

6
src/llmtuner/hparams/parser.py
View File

@ -174,9 +174,6 @@ def get_train_args(args: Optional[Dict[str, Any]] = None) -> _TRAIN_CLS:
):
raise ValueError("Distributed training does not support layer-wise GaLore.")
if finetuning_args.use_galore and training_args.deepspeed is not None:
raise ValueError("GaLore is incompatible with DeepSpeed yet.")
if (
finetuning_args.use_badam
and finetuning_args.badam_mode == "layer"
@ -184,6 +181,9 @@ def get_train_args(args: Optional[Dict[str, Any]] = None) -> _TRAIN_CLS:
):
raise ValueError("Layer-wise BAdam does not yet support distributed training, use ratio-wise BAdam.")
if (finetuning_args.use_galore or finetuning_args.use_badam) and training_args.deepspeed is not None:
raise ValueError("GaLore and BAdam are incompatible with DeepSpeed yet.")
if model_args.infer_backend == "vllm":
raise ValueError("vLLM backend is only available for API, CLI and Web.")