namespace-Pt commited on
Commit
1271e80
1 Parent(s): 08bdd0b

Upload download.py with huggingface_hub

Browse files
Files changed (1) hide show
  1. download.py +71 -0
download.py ADDED
@@ -0,0 +1,71 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ import os
2
+ import glob
3
+ from datasets import load_dataset, get_dataset_config_names
4
+ from dataclasses import dataclass, field
5
+ from transformers import AutoModel, AutoTokenizer, AutoModelForCausalLM, AutoModelForSeq2SeqLM, HfArgumentParser
6
+ from typing import Optional, List
7
+
8
+
9
+ @dataclass
10
+ class DownloadArgs:
11
+ model_cache_dir: str = field(
12
+ default='/share/LMs',
13
+ metadata={'help': 'Default path to save language models'}
14
+ )
15
+ model_name_or_path: Optional[str] = field(
16
+ default=None,
17
+ metadata={'help': 'Path to pretrained model or model identifier from huggingface.co/models'}
18
+ )
19
+ use_lm_class: bool = field(
20
+ default=False,
21
+ metadata={'help': 'Call .from_pretrained from AutoModelForCausalLM? Useful when downloading remote-code based lms.'}
22
+ )
23
+ dataset_cache_dir: str = field(
24
+ default='/share/peitian/Data/Datasets/huggingface',
25
+ metadata={'help': 'Default path to save huggingface datasets'}
26
+ )
27
+ dataset_name_or_path: Optional[str] = field(
28
+ default=None,
29
+ metadata={'help': 'Dataset name'}
30
+ )
31
+ dataset_subset: Optional[str] = field(
32
+ default=None,
33
+ metadata={'help': 'Dataset subset name'}
34
+ )
35
+ dataset_split: Optional[str] = field(
36
+ default=None,
37
+ metadata={'help': 'Dataset split'}
38
+ )
39
+
40
+ revision: str = field(
41
+ default=None,
42
+ metadata={'help': 'Remote code revision'}
43
+ )
44
+ resume_download: bool = field(
45
+ default=False,
46
+ metadata={'help': 'Resume downloading'}
47
+ )
48
+ def __post_init__(self):
49
+ # folder or model not exists
50
+ if self.model_name_or_path is not None:
51
+ kwargs = {
52
+ 'revision': self.revision,
53
+ 'resume_download': self.resume_download
54
+ }
55
+ AutoTokenizer.from_pretrained(self.model_name_or_path, cache_dir=self.model_cache_dir, trust_remote_code=True, **kwargs)
56
+ if self.use_lm_class:
57
+ AutoModelForCausalLM.from_pretrained(self.model_name_or_path, cache_dir=self.model_cache_dir, trust_remote_code=True, **kwargs)
58
+ else:
59
+ AutoModel.from_pretrained(self.model_name_or_path, cache_dir=self.model_cache_dir, trust_remote_code=True, **kwargs)
60
+ if self.dataset_name_or_path is not None:
61
+ if self.dataset_subset is None:
62
+ dataset_subsets = get_dataset_config_names(self.dataset_name_or_path)
63
+ for dataset_subset in dataset_subsets:
64
+ load_dataset(self.dataset_name_or_path, name=dataset_subset, split=self.dataset_split, cache_dir=self.dataset_cache_dir)
65
+ else:
66
+ load_dataset(self.dataset_name_or_path, name=self.dataset_subset, split=self.dataset_split, cache_dir=self.dataset_cache_dir)
67
+
68
+
69
+ if __name__ == "__main__":
70
+ parser = HfArgumentParser([DownloadArgs])
71
+ args, = parser.parse_args_into_dataclasses()