mirror of
https://github.com/hiyouga/LLaMA-Factory.git
synced 2025-08-01 11:12:50 +08:00
tiny fix
Former-commit-id: 44298c12355082740857ba650bf44a18d4d3b40d
This commit is contained in:
parent
b982a9df83
commit
e94cf814ff
@ -21,13 +21,13 @@
|
||||
## Supported Training Approaches
|
||||
|
||||
- [(Continually) pre-training](https://s3-us-west-2.amazonaws.com/openai-assets/research-covers/language-unsupervised/language_understanding_paper.pdf)
|
||||
- Full-parameter training
|
||||
- Partial-parameter training
|
||||
- Full-parameter tuning
|
||||
- Partial-parameter tuning
|
||||
- [LoRA](https://arxiv.org/abs/2106.09685)
|
||||
- [QLoRA](https://arxiv.org/abs/2305.14314)
|
||||
- [Supervised fine-tuning](https://arxiv.org/abs/2109.01652)
|
||||
- Full-parameter training
|
||||
- Partial-parameter training
|
||||
- Full-parameter tuning
|
||||
- Partial-parameter tuning
|
||||
- [LoRA](https://arxiv.org/abs/2106.09685)
|
||||
- [QLoRA](https://arxiv.org/abs/2305.14314)
|
||||
- [RLHF](https://arxiv.org/abs/2203.02155)
|
||||
|
@ -261,8 +261,8 @@ def prepare_args(
|
||||
if training_args.do_predict and (not training_args.predict_with_generate):
|
||||
raise ValueError("Please enable `predict_with_generate` to save model predictions.")
|
||||
|
||||
if model_args.quantization_bit is not None and finetuning_args.finetuning_type != "lora":
|
||||
raise ValueError("Quantization is only compatible with the LoRA method.")
|
||||
if model_args.quantization_bit is not None and finetuning_args.finetuning_type == "full":
|
||||
raise ValueError("Quantization is incompatible with the full-parameter tuning.")
|
||||
|
||||
if model_args.quantization_bit is not None and (not training_args.do_train):
|
||||
logger.warning("Evaluating model in 4/8-bit mode may cause lower scores.")
|
||||
|
@ -148,7 +148,8 @@ class DataTrainingArguments:
|
||||
|
||||
def __post_init__(self): # support mixing multiple datasets
|
||||
dataset_names = [ds.strip() for ds in self.dataset.split(",")]
|
||||
dataset_info = json.load(open(os.path.join(self.dataset_dir, "dataset_info.json"), "r"))
|
||||
with open(os.path.join(self.dataset_dir, "dataset_info.json"), "r") as f:
|
||||
dataset_info = json.load(f)
|
||||
|
||||
self.dataset_list: List[DatasetAttr] = []
|
||||
for name in dataset_names:
|
||||
|
Loading…
x
Reference in New Issue
Block a user