bring back images
Browse files- SNLI-VE.py +7 -13
SNLI-VE.py
CHANGED
@@ -62,10 +62,11 @@ _SNLI_VE_SPLITS = {
|
|
62 |
"validation": "snli_ve_dev.jsonl",
|
63 |
"test": "snli_ve_test.jsonl",
|
64 |
}
|
65 |
-
|
66 |
|
67 |
_FEATURES = datasets.Features(
|
68 |
-
{
|
|
|
69 |
"filename": datasets.Value("string"),
|
70 |
"premise": datasets.Value("string"),
|
71 |
"hypothesis": datasets.Value("string"),
|
@@ -77,15 +78,7 @@ _FEATURES = datasets.Features(
|
|
77 |
class SNLIVE(datasets.GeneratorBasedBuilder):
|
78 |
"""SNLIVE."""
|
79 |
|
80 |
-
|
81 |
-
# def manual_download_instructions(self):
|
82 |
-
# return """\
|
83 |
-
# In order to get the flickr data on which SNLI-VE is built, You need to go to http://shannon.cs.illinois.edu/DenotationGraph/data/index.html,
|
84 |
-
# and manually download the dataset ("Flickr 30k images."). Once it is completed,
|
85 |
-
# a file named `flickr30k-images.tar.gz` will appear in your Downloads folder
|
86 |
-
# or whichever folder your browser chooses to save files to.
|
87 |
-
# Then, the dataset can be loaded using the following command `datasets.load_dataset("flickr30k", data_dir="<path/to/folder>")`.
|
88 |
-
# """
|
89 |
DEFAULT_CONFIG_NAME = "Default"
|
90 |
logger.warning("HER0")
|
91 |
def _info(self):
|
@@ -108,7 +101,7 @@ class SNLIVE(datasets.GeneratorBasedBuilder):
|
|
108 |
},
|
109 |
}
|
110 |
snli_ve_annotation_path = dl_manager.download_and_extract(urls)
|
111 |
-
images_path =
|
112 |
|
113 |
return [
|
114 |
datasets.SplitGenerator(
|
@@ -142,8 +135,9 @@ class SNLIVE(datasets.GeneratorBasedBuilder):
|
|
142 |
for elem in json_file:
|
143 |
elem = json.loads(elem)
|
144 |
img_filename = str(elem["Flickr30K_ID"]) + ".jpg"
|
145 |
-
|
146 |
record = {
|
|
|
147 |
"filename": img_filename,
|
148 |
"premise": elem["sentence1"],
|
149 |
"hypothesis": elem["sentence2"],
|
|
|
62 |
"validation": "snli_ve_dev.jsonl",
|
63 |
"test": "snli_ve_test.jsonl",
|
64 |
}
|
65 |
+
JZ_FOLDER_PATH = f"{os.environ['cnw_ALL_CCFRSCRATCH']}/local_datasets/flickr30k-images.tar.gz"
|
66 |
|
67 |
_FEATURES = datasets.Features(
|
68 |
+
{
|
69 |
+
"image": datasets.Image(),
|
70 |
"filename": datasets.Value("string"),
|
71 |
"premise": datasets.Value("string"),
|
72 |
"hypothesis": datasets.Value("string"),
|
|
|
78 |
class SNLIVE(datasets.GeneratorBasedBuilder):
|
79 |
"""SNLIVE."""
|
80 |
|
81 |
+
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
82 |
DEFAULT_CONFIG_NAME = "Default"
|
83 |
logger.warning("HER0")
|
84 |
def _info(self):
|
|
|
101 |
},
|
102 |
}
|
103 |
snli_ve_annotation_path = dl_manager.download_and_extract(urls)
|
104 |
+
images_path = dl_manager.download_and_extract(JZ_FOLDER_PATH)
|
105 |
|
106 |
return [
|
107 |
datasets.SplitGenerator(
|
|
|
135 |
for elem in json_file:
|
136 |
elem = json.loads(elem)
|
137 |
img_filename = str(elem["Flickr30K_ID"]) + ".jpg"
|
138 |
+
assert os.path.exists(os.path.join(images_path, img_filename))
|
139 |
record = {
|
140 |
+
"image": img_filename,
|
141 |
"filename": img_filename,
|
142 |
"premise": elem["sentence1"],
|
143 |
"hypothesis": elem["sentence2"],
|