File size: 1,864 Bytes
96fe658 |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 |
import os
os.environ['CUDA_VISIBLE_DEVICES'] = '0'
kwargs = {
'per_device_train_batch_size': 2,
'per_device_eval_batch_size': 2,
'save_steps': 50,
'gradient_accumulation_steps': 4,
'num_train_epochs': 1,
}
def test_llm():
from swift.llm import TrainArguments, sft_main, infer_main, InferArguments
result = sft_main(
TrainArguments(
model='Qwen/Qwen2.5-1.5B-Instruct',
train_type='lora',
num_labels=2,
dataset=['DAMO_NLP/jd:cls#2000'],
split_dataset_ratio=0.01,
**kwargs))
last_model_checkpoint = result['last_model_checkpoint']
infer_main(InferArguments(adapters=last_model_checkpoint, load_data_args=True))
def test_bert():
from swift.llm import TrainArguments, sft_main, infer_main, InferArguments
result = sft_main(
TrainArguments(
model='answerdotai/ModernBERT-base',
# model='iic/nlp_structbert_backbone_base_std',
train_type='full',
num_labels=2,
dataset=['DAMO_NLP/jd:cls#2000'],
split_dataset_ratio=0.01,
**kwargs))
last_model_checkpoint = result['last_model_checkpoint']
infer_main(InferArguments(model=last_model_checkpoint, load_data_args=True))
def test_mllm():
from swift.llm import TrainArguments, sft_main, infer_main, InferArguments
result = sft_main(
TrainArguments(
model='OpenGVLab/InternVL2-1B',
train_type='lora',
num_labels=2,
dataset=['DAMO_NLP/jd:cls#500'],
split_dataset_ratio=0.01,
**kwargs))
last_model_checkpoint = result['last_model_checkpoint']
infer_main(InferArguments(adapters=last_model_checkpoint, load_data_args=True))
if __name__ == '__main__':
# test_llm()
# test_bert()
test_mllm()
|