Skip to content

use max_length to replace max_seq_length; correct README for #2862

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
merged 4 commits into from
Oct 23, 2025
+15 −15
Merged

use max_length to replace max_seq_length; correct README for #2862

Show file tree
Hide file tree
Changes from 3 commits
Commits
Show all changes
4 commits
Select commit Hold shift + click to select a range
7a7fb87
use `max_length` to replace `max_seq_length`; correct README for
kaixuanliu Oct 23, 2025
2e5e99a
Revert "use `max_length` to replace `max_seq_length`; correct README …
kaixuanliu Oct 23, 2025
b87e092
use `max_length` for examples
kaixuanliu Oct 23, 2025
6a3909a
revert the change in sft
kaixuanliu Oct 23, 2025
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
2 changes: 1 addition & 1 deletion examples/alora_finetuning/README.md
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -32,7 +32,7 @@ trainer = Trainer(
model=peft_model,
train_dataset=dataset,
dataset_text_field="text",
max_seq_length=2048,
max_length=2048,
tokenizer=tokenizer,
data_collator=data_collator,
)
Expand Down
2 changes: 1 addition & 1 deletion examples/bone_finetuning/README.md
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -28,7 +28,7 @@ peft_model.print_trainable_parameters()

dataset = load_dataset("imdb", split="train[:1%]")

training_args = SFTConfig(dataset_text_field="text", max_seq_length=128)
training_args = SFTConfig(dataset_text_field="text", max_length=128)
trainer = SFTTrainer(
model=peft_model,
args=training_args,
Expand Down
2 changes: 1 addition & 1 deletion examples/corda_finetuning/README.md
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -109,7 +109,7 @@ preprocess_corda(model, lora_config, run_model=run_model)
peft_model = get_peft_model(model, lora_config)
peft_model.print_trainable_parameters()

training_args = SFTConfig(dataset_text_field="text", max_seq_length=128)
training_args = SFTConfig(dataset_text_field="text", max_length=128)
trainer = SFTTrainer(
model=peft_model,
args=training_args,
Expand Down
6 changes: 3 additions & 3 deletions examples/delora_finetuning/README.md
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -26,7 +26,7 @@ peft_model.print_trainable_parameters()

dataset = load_dataset("imdb", split="train[:1%]")

training_args = SFTConfig(dataset_text_field="text", max_seq_length=128)
training_args = SFTConfig(dataset_text_field="text", max_length=128)
trainer = SFTTrainer(
model=peft_model,
args=training_args,
Expand All @@ -52,7 +52,7 @@ peft_model = PeftModel.from_pretrained(model, "delora-llama-3-8b")
## Advanced Usage
In this script the default DeLoRA layers are the query and value layers of the Llama model. Adding adapters on more layers will increase memory usage. If you wish to choose a different set of layers for DeLoRA to be applied on, you can simply define it using:
```bash
python examples/delora_finetuning/delora_finetuning.py --base_model meta-llama/Meta-Llama-3-8B --delora_target_modules "q_proj,k_proj,v_proj,o_proj"
python examples/delora_finetuning/delora_finetuning.py --base_model meta-llama/Meta-Llama-3-8B --target_modules "q_proj,k_proj,v_proj,o_proj"
```

Using different lambdas for different layers is also possible by setting `lambda_pattern`.
Expand All @@ -74,7 +74,7 @@ python delora_finetuning.py \
--rank 32 \
--delora_lambda 15 \
--module_dropout 0.1 \
--delora_target_modules "q_proj,v_proj" \
--target_modules "q_proj,v_proj" \
--hub_model_id "YOUR_HF_REPO" \
--push_to_hub
```
Expand Down
2 changes: 1 addition & 1 deletion examples/dora_finetuning/README.md
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -24,7 +24,7 @@ trainer = transformers.Trainer(
model=peft_model,
train_dataset=dataset,
dataset_text_field="text",
max_seq_length=2048,
max_length=2048,
tokenizer=tokenizer,
)
trainer.train()
Expand Down
2 changes: 1 addition & 1 deletion examples/lorafa_finetune/README.md
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -40,7 +40,7 @@ trainer = transformers.Trainer(
model=peft_model,
train_dataset=dataset,
dataset_text_field="text",
max_seq_length=2048,
max_length=2048,
processing_class=tokenizer,
optimizers=(optimizer, None),
)
Expand Down
2 changes: 1 addition & 1 deletion examples/miss_finetuning/README.md
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -36,7 +36,7 @@ peft_model.print_trainable_parameters()

dataset = load_dataset("imdb", split="train[:1%]")

training_args = SFTConfig(dataset_text_field="text", max_seq_length=128)
training_args = SFTConfig(dataset_text_field="text", max_length=128)
trainer = SFTTrainer(
model=peft_model,
args=training_args,
Expand Down
2 changes: 1 addition & 1 deletion examples/olora_finetuning/README.md
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -18,7 +18,7 @@ lora_config = LoraConfig(
init_lora_weights="olora"
)
peft_model = get_peft_model(model, lora_config)
training_args = SFTConfig(dataset_text_field="text", max_seq_length=128)
training_args = SFTConfig(dataset_text_field="text", max_length=128)
trainer = SFTTrainer(
model=peft_model,
train_dataset=dataset,
Expand Down
2 changes: 1 addition & 1 deletion examples/pissa_finetuning/README.md
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -23,7 +23,7 @@ peft_model.print_trainable_parameters()

dataset = load_dataset("imdb", split="train[:1%]")

training_args = SFTConfig(dataset_text_field="text", max_seq_length=128)
training_args = SFTConfig(dataset_text_field="text", max_length=128)
trainer = SFTTrainer(
model=peft_model,
args=training_args,
Expand Down
2 changes: 1 addition & 1 deletion examples/randlora_finetuning/README.md
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -20,7 +20,7 @@ trainer = transformers.Trainer(
model=peft_model,
train_dataset=dataset,
dataset_text_field="text",
max_seq_length=2048,
max_length=2048,
processing_class=tokenizer,
)
trainer.train()
Expand Down
2 changes: 1 addition & 1 deletion examples/road_finetuning/README.md
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -26,7 +26,7 @@ trainer = transformers.Trainer(
model=peft_model,
train_dataset=dataset,
dataset_text_field="text",
max_seq_length=2048,
max_length=2048,
tokenizer=tokenizer,
)
trainer.train()
Expand Down
2 changes: 1 addition & 1 deletion examples/sft/train.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -18,7 +18,7 @@ class ModelArguments:
model_name_or_path: str = field(
metadata={"help": "Path to pretrained model or model identifier from huggingface.co/models"}
)
max_seq_length: Optional[int] = field(
max_length: Optional[int] = field(
Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

This change isn't enough, because there is a conflicting argument with the same name. I already took care of that in #2857, so you can revert this change here.

Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Oops, I changed the code here by mistake. Yes, for sft case, we should take more actions. Have reverted.

default=512,
metadata={"help": "The maximum total input sequence length after tokenization."},
)
Expand Down
2 changes: 1 addition & 1 deletion examples/shira_finetuning/README.md
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -18,7 +18,7 @@ shira_config = ShiraConfig(
r=32,
)
peft_model = get_peft_model(model, shira_config)
training_args = SFTConfig(dataset_text_field="text", max_seq_length=128)
training_args = SFTConfig(dataset_text_field="text", max_length=128)
trainer = SFTTrainer(
model=peft_model,
train_dataset=dataset,
Expand Down
2 changes: 1 addition & 1 deletion examples/waveft_finetuning/README.md
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -20,7 +20,7 @@ waveft_config = WaveFTConfig(
n_frequency=2592,
)
peft_model = get_peft_model(model, waveft_config)
training_args = SFTConfig(dataset_text_field="text", max_seq_length=128)
training_args = SFTConfig(dataset_text_field="text", max_length=128)
trainer = SFTTrainer(
model=peft_model,
train_dataset=dataset,
Expand Down