ros/LLaMA-Factory
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

fix tests

Browse Source 
Former-commit-id: 23f97bd437424ef43b2b84743d56acc5d1ca70d5
This commit is contained in:
hiyouga
2024-01-20 19:58:04 +08:00
parent 80637fc06d
commit 1750218057
12 changed files with 80 additions and 65 deletions
Download Patch File Download Diff File Expand all files Collapse all files

2
src/llmtuner/train/dpo/workflow.py
View File

@@ -25,7 +25,7 @@ def run_dpo(
callbacks: Optional[List["TrainerCallback"]] = None
):
model, tokenizer = load_model_and_tokenizer(model_args, finetuning_args, training_args.do_train)
dataset = get_dataset(model_args, data_args, tokenizer, training_args, stage="rm")
dataset = get_dataset(tokenizer, model_args, data_args, training_args, stage="rm")
data_collator = DPODataCollatorWithPadding(
tokenizer=tokenizer,
pad_to_multiple_of=8,

2
src/llmtuner/train/ppo/workflow.py
View File

@@ -29,7 +29,7 @@ def run_ppo(
callbacks: Optional[List["TrainerCallback"]] = None
):
model, tokenizer = load_model_and_tokenizer(model_args, finetuning_args, training_args.do_train, add_valuehead=True)
dataset = get_dataset(model_args, data_args, tokenizer, training_args, stage="ppo")
dataset = get_dataset(tokenizer, model_args, data_args, training_args, stage="ppo")
tokenizer.padding_side = "left" # use left-padding in generation while using right-padding in training
data_collator = DataCollatorWithPadding(tokenizer=tokenizer)

2
src/llmtuner/train/pt/workflow.py
View File

@@ -22,7 +22,7 @@ def run_pt(
callbacks: Optional[List["TrainerCallback"]] = None
):
model, tokenizer = load_model_and_tokenizer(model_args, finetuning_args, training_args.do_train)
dataset = get_dataset(model_args, data_args, tokenizer, training_args, stage="pt")
dataset = get_dataset(tokenizer, model_args, data_args, training_args, stage="pt")
data_collator = DataCollatorForLanguageModeling(tokenizer=tokenizer, mlm=False)
# Initialize our Trainer

2
src/llmtuner/train/rm/workflow.py
View File

@@ -26,7 +26,7 @@ def run_rm(
callbacks: Optional[List["TrainerCallback"]] = None
):
model, tokenizer = load_model_and_tokenizer(model_args, finetuning_args, training_args.do_train, add_valuehead=True)
dataset = get_dataset(model_args, data_args, tokenizer, training_args, stage="rm")
dataset = get_dataset(tokenizer, model_args, data_args, training_args, stage="rm")
data_collator = PairwiseDataCollatorWithPadding(tokenizer, pad_to_multiple_of=8)
# Update arguments

2
src/llmtuner/train/sft/workflow.py
View File

@@ -27,7 +27,7 @@ def run_sft(
callbacks: Optional[List["TrainerCallback"]] = None
):
model, tokenizer = load_model_and_tokenizer(model_args, finetuning_args, training_args.do_train)
dataset = get_dataset(model_args, data_args, tokenizer, training_args, stage="sft")
dataset = get_dataset(tokenizer, model_args, data_args, training_args, stage="sft")
if training_args.predict_with_generate:
tokenizer.padding_side = "left" # use left-padding in generation