SUMIN/HuggingFace_transformer
1
0
Fork 0
Code Issues Pull Requests Actions 7 Packages Projects Releases Wiki Activity

Simplify and update trl examples (#38772)

Browse Source 
* Simplify and update trl examples

* Remove optim_args from SFTConfig in Trainer documentation

* Update docs/source/en/trainer.md

* Apply suggestions from code review

* Update docs/source/en/trainer.md

Co-authored-by: Steven Liu <59462357+stevhliu@users.noreply.github.com>

---------

Co-authored-by: Quentin Gallouédec <qgallouedec@Quentins-MacBook-Pro.local>
Co-authored-by: Steven Liu <59462357+stevhliu@users.noreply.github.com>
This commit is contained in:
Quentin Gallouédec
2025-06-13 14:03:49 +02:00
committed by GitHub
parent de24fb63ed
commit c989ddd294
7 changed files with 114 additions and 347 deletions
Download Patch File Download Diff File Expand all files Collapse all files

28
docs/source/ko/model_doc/mamba.md
View File

@@ -58,34 +58,18 @@ print(tokenizer.batch_decode(out))
```python
from datasets import load_dataset
from trl import SFTTrainer
from trl import SFTConfig, SFTTrainer
from peft import LoraConfig
from transformers import AutoTokenizer, AutoModelForCausalLM, TrainingArguments
model_id = "state-spaces/mamba-130m-hf"
tokenizer = AutoTokenizer.from_pretrained(model_id)
model = AutoModelForCausalLM.from_pretrained(model_id)
dataset = load_dataset("Abirate/english_quotes", split="train")
training_args = TrainingArguments(
output_dir="./results",
num_train_epochs=3,
per_device_train_batch_size=4,
logging_dir='./logs',
logging_steps=10,
learning_rate=2e-3
)
lora_config = LoraConfig(
r=8,
target_modules=["x_proj", "embeddings", "in_proj", "out_proj"],
task_type="CAUSAL_LM",
bias="none"
)
training_args = SFTConfig(dataset_text_field="quote")
lora_config = LoraConfig(target_modules=["x_proj", "embeddings", "in_proj", "out_proj"])
trainer = SFTTrainer(
model=model,
tokenizer=tokenizer,
model=model_id,
args=training_args,
peft_config=lora_config,
train_dataset=dataset,
dataset_text_field="quote",
peft_config=lora_config,
)
trainer.train()
```

35
docs/source/ko/model_doc/mamba2.md
View File

@@ -57,40 +57,19 @@ print(tokenizer.batch_decode(out))
이곳은 미세조정을 위한 초안 스크립트입니다:
```python
from trl import SFTTrainer
from datasets import load_dataset
from peft import LoraConfig
from transformers import AutoTokenizer, Mamba2ForCausalLM, TrainingArguments
model_id = 'mistralai/Mamba-Codestral-7B-v0.1'
tokenizer = AutoTokenizer.from_pretrained(model_id, revision='refs/pr/9', from_slow=True, legacy=False)
tokenizer.pad_token = tokenizer.eos_token
tokenizer.padding_side = "left" #왼쪽 패딩으로 설정
from trl import SFTConfig, SFTTrainer
model = Mamba2ForCausalLM.from_pretrained(model_id, revision='refs/pr/9')
model_id = "mistralai/Mamba-Codestral-7B-v0.1"
dataset = load_dataset("Abirate/english_quotes", split="train")
# CUDA 커널없이는, 배치크기 2가 80GB 장치를 하나 차지합니다.
# 하지만 정확도는 감소합니다.
# 실험과 시도를 환영합니다!
training_args = TrainingArguments(
output_dir="./results",
num_train_epochs=3,
per_device_train_batch_size=2,
logging_dir='./logs',
logging_steps=10,
learning_rate=2e-3
)
lora_config = LoraConfig(
r=8,
target_modules=["embeddings", "in_proj", "out_proj"],
task_type="CAUSAL_LM",
bias="none"
)
training_args = SFTConfig(dataset_text_field="quote", gradient_checkpointing=True, per_device_train_batch_size=4)
lora_config = LoraConfig(target_modules=["x_proj", "embeddings", "in_proj", "out_proj"])
trainer = SFTTrainer(
model=model,
tokenizer=tokenizer,
model=model_id,
args=training_args,
peft_config=lora_config,
train_dataset=dataset,
dataset_text_field="quote",
peft_config=lora_config,
)
trainer.train()
```