Create xtd_11.py
Browse files
xtd_11.py
ADDED
@@ -0,0 +1,86 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
from datasets import DatasetDict, load_dataset, SplitGenerator, Split
|
2 |
+
import datasets
|
3 |
+
|
4 |
+
_Base_url = "https://huggingface.co/datasets/khalidalt/xtd_11/resolve/main/test/"
|
5 |
+
_Languages = ["ar", "de", "en", "es", "fr", "jp", "ko", "pl", "ru", "tr", "zh"] # List all your languages here
|
6 |
+
|
7 |
+
class XTD_11Config(datasets.BuilderConfig):
|
8 |
+
""" Builder config for Joud Dataset. """
|
9 |
+
|
10 |
+
def __init__(self, subset, **kwargs):
|
11 |
+
super(JoudConfig, self).__init__(**kwargs)
|
12 |
+
|
13 |
+
if subset !="all":
|
14 |
+
|
15 |
+
self.subset = [subset]
|
16 |
+
else:
|
17 |
+
self.subset = _SUBSETS
|
18 |
+
|
19 |
+
|
20 |
+
class XTD_11(datasets.GeneratorBasedBuilder):
|
21 |
+
VERSION = datasets.Version("1.0.0")
|
22 |
+
BUILDER_CONFIGS_CLASS = XTD_11Config
|
23 |
+
BUILDER_CONFIGS = [
|
24 |
+
XTD_11Config(name=subset,
|
25 |
+
subset=subset,
|
26 |
+
version=datasets.Version("1.1.0", ""),
|
27 |
+
description='')
|
28 |
+
for subset in _SUBSETS
|
29 |
+
] + [
|
30 |
+
XTD_11Config(
|
31 |
+
name = "all",
|
32 |
+
subset="all",
|
33 |
+
version=datasets.Version("1.1.0", ""),
|
34 |
+
description='')
|
35 |
+
]
|
36 |
+
def _info(self):
|
37 |
+
# Assumes all files have the same structure. Adjust as needed.
|
38 |
+
return datasets.DatasetInfo(
|
39 |
+
# This is a placeholder, adjust according to your dataset structure
|
40 |
+
features=datasets.Features({
|
41 |
+
"text": datasets.Value("string"),
|
42 |
+
"image_name": dataset.Value("string"),
|
43 |
+
"url": dataset.Value("string"),
|
44 |
+
}),
|
45 |
+
description="A benchmark to test model capability in image retrieval named xtd.",
|
46 |
+
)
|
47 |
+
|
48 |
+
def _split_generators(self, dl_manager):
|
49 |
+
"""Returns SplitGenerators."""
|
50 |
+
# Define your base URL or local path where the files are located
|
51 |
+
|
52 |
+
# Generate splits for each language
|
53 |
+
splits = []
|
54 |
+
for lang in _Languages:
|
55 |
+
splits.append(SplitGenerator(
|
56 |
+
name=lang,
|
57 |
+
# Assuming the use of URLs. For local files, adjust accordingly.
|
58 |
+
gen_kwargs={"filepath": f"{_Base_url}{lang}.json"}
|
59 |
+
))
|
60 |
+
|
61 |
+
# Optionally, define a combined split that includes all languages
|
62 |
+
# This could be more complex to implement, depending on how you want to combine the data
|
63 |
+
# Here's a conceptual placeholder for the combined split
|
64 |
+
splits.append(SplitGenerator(
|
65 |
+
name="combined_test",
|
66 |
+
gen_kwargs={"filepaths": [f"{_Base_url}{lang}.json" for lang in languages]}
|
67 |
+
))
|
68 |
+
|
69 |
+
return splits
|
70 |
+
|
71 |
+
def _generate_examples(self, filepath=None, filepaths=None):
|
72 |
+
"""Yields examples."""
|
73 |
+
# This method needs to handle both single and multiple filepaths
|
74 |
+
# Adjust the logic based on how you want to load and yield data from your JSON files
|
75 |
+
|
76 |
+
if filepath:
|
77 |
+
# Load data from a single file
|
78 |
+
with open(filepath, encoding="utf-8") as f:
|
79 |
+
for id, row in enumerate(json.load(f)):
|
80 |
+
yield id, row
|
81 |
+
elif filepaths:
|
82 |
+
# Example logic for combined split, adjust as needed
|
83 |
+
for fp in filepaths:
|
84 |
+
with open(fp, encoding="utf-8") as f:
|
85 |
+
for id, row in enumerate(json.load(f)):
|
86 |
+
yield id, row
|