File size: 1,192 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
import os

os.environ['CUDA_VISIBLE_DEVICES'] = '0'

kwargs = {
    'per_device_train_batch_size': 2,
    'save_steps': 5,
    'gradient_accumulation_steps': 4,
    'num_train_epochs': 1,
}


def test_llm():
    from swift.llm import pt_main, TrainArguments, infer_main, InferArguments
    result = pt_main(
        TrainArguments(
            model='Qwen/Qwen2-7B-Instruct', dataset=['swift/sharegpt:all#100'], split_dataset_ratio=0.01, **kwargs))
    last_model_checkpoint = result['last_model_checkpoint']
    infer_main(InferArguments(adapters=last_model_checkpoint, load_data_args=True, merge_lora=True))


def test_mllm():
    from swift.llm import pt_main, TrainArguments, infer_main, InferArguments
    result = pt_main(
        TrainArguments(
            model='Qwen/Qwen2-VL-7B-Instruct',
            dataset=['modelscope/coco_2014_caption:validation#20', 'AI-ModelScope/alpaca-gpt4-data-en#20'],
            split_dataset_ratio=0.01,
            **kwargs))
    last_model_checkpoint = result['last_model_checkpoint']
    infer_main(InferArguments(adapters=last_model_checkpoint, load_data_args=True, merge_lora=True))


if __name__ == '__main__':
    # test_llm()
    test_mllm()