Jinchen Ge
commited on
Commit
•
62c0a9b
1
Parent(s):
2bb73b8
Add test set
Browse files- vqa-lxmert.py +46 -23
vqa-lxmert.py
CHANGED
@@ -47,11 +47,14 @@ _URLS = {
|
|
47 |
"train_feat": "https://nlp.cs.unc.edu/data/lxmert_data/mscoco_imgfeat/train2014_obj36.zip",
|
48 |
"valid": "https://nlp.cs.unc.edu/data/lxmert_data/vqa/valid.json",
|
49 |
"valid_feat": "https://nlp.cs.unc.edu/data/lxmert_data/mscoco_imgfeat/val2014_obj36.zip",
|
|
|
|
|
50 |
"ans2label": "https://raw.githubusercontent.com/airsplay/lxmert/master/data/vqa/trainval_ans2label.json",
|
51 |
}
|
52 |
|
53 |
-
|
54 |
-
|
|
|
55 |
|
56 |
FIELDNAMES = [
|
57 |
"img_id", "img_h", "img_w", "objects_id", "objects_conf", "attrs_id", "attrs_conf", "num_boxes", "boxes", "features"
|
@@ -102,11 +105,15 @@ class VqaV2Lxmert(datasets.GeneratorBasedBuilder):
|
|
102 |
return [
|
103 |
datasets.SplitGenerator(
|
104 |
name=datasets.Split.TRAIN,
|
105 |
-
gen_kwargs={"filepath": dl_dir["train"], "imgfeat": os.path.join(dl_dir["train_feat"],
|
106 |
),
|
107 |
datasets.SplitGenerator(
|
108 |
name=datasets.Split.VALIDATION,
|
109 |
-
gen_kwargs={"filepath": dl_dir["valid"], "imgfeat": os.path.join(dl_dir["valid_feat"],
|
|
|
|
|
|
|
|
|
110 |
),
|
111 |
]
|
112 |
|
@@ -135,26 +142,42 @@ class VqaV2Lxmert(datasets.GeneratorBasedBuilder):
|
|
135 |
normalized_boxes[:, (1, 3)] /= img_h
|
136 |
return normalized_boxes
|
137 |
|
138 |
-
def _generate_examples(self, filepath, imgfeat):
|
139 |
""" Yields examples as (key, example) tuples."""
|
140 |
id2features = self._load_features(imgfeat)
|
141 |
with open(filepath, encoding="utf-8") as f:
|
142 |
vqa = json.load(f)
|
143 |
-
|
144 |
-
|
145 |
-
|
146 |
-
|
147 |
-
|
148 |
-
|
149 |
-
|
150 |
-
|
151 |
-
|
152 |
-
|
153 |
-
|
154 |
-
|
155 |
-
|
156 |
-
"
|
157 |
-
|
158 |
-
|
159 |
-
|
160 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
47 |
"train_feat": "https://nlp.cs.unc.edu/data/lxmert_data/mscoco_imgfeat/train2014_obj36.zip",
|
48 |
"valid": "https://nlp.cs.unc.edu/data/lxmert_data/vqa/valid.json",
|
49 |
"valid_feat": "https://nlp.cs.unc.edu/data/lxmert_data/mscoco_imgfeat/val2014_obj36.zip",
|
50 |
+
"test": "https://nlp.cs.unc.edu/data/lxmert_data/vqa/test.json",
|
51 |
+
"test_feat": "https://nlp.cs.unc.edu/data/lxmert_data/mscoco_imgfeat/test2015_obj36.zip",
|
52 |
"ans2label": "https://raw.githubusercontent.com/airsplay/lxmert/master/data/vqa/trainval_ans2label.json",
|
53 |
}
|
54 |
|
55 |
+
_TRAIN_FEAT_PATH = "train2014_obj36.tsv"
|
56 |
+
_VALID_FEAT_PATH = "mscoco_imgfeat/val2014_obj36.tsv"
|
57 |
+
_TEST_FEAT_PATH = "mscoco_imgfeat/test2015_obj36.tsv"
|
58 |
|
59 |
FIELDNAMES = [
|
60 |
"img_id", "img_h", "img_w", "objects_id", "objects_conf", "attrs_id", "attrs_conf", "num_boxes", "boxes", "features"
|
|
|
105 |
return [
|
106 |
datasets.SplitGenerator(
|
107 |
name=datasets.Split.TRAIN,
|
108 |
+
gen_kwargs={"filepath": dl_dir["train"], "imgfeat": os.path.join(dl_dir["train_feat"], _TRAIN_FEAT_PATH)},
|
109 |
),
|
110 |
datasets.SplitGenerator(
|
111 |
name=datasets.Split.VALIDATION,
|
112 |
+
gen_kwargs={"filepath": dl_dir["valid"], "imgfeat": os.path.join(dl_dir["valid_feat"], _VALID_FEAT_PATH)},
|
113 |
+
),
|
114 |
+
datasets.SplitGenerator(
|
115 |
+
name=datasets.Split.TEST,
|
116 |
+
gen_kwargs={"filepath": dl_dir["test"], "imgfeat": os.path.join(dl_dir["test_feat"], _TEST_FEAT_PATH), "labeled": False},
|
117 |
),
|
118 |
]
|
119 |
|
|
|
142 |
normalized_boxes[:, (1, 3)] /= img_h
|
143 |
return normalized_boxes
|
144 |
|
145 |
+
def _generate_examples(self, filepath, imgfeat, labeled=True):
|
146 |
""" Yields examples as (key, example) tuples."""
|
147 |
id2features = self._load_features(imgfeat)
|
148 |
with open(filepath, encoding="utf-8") as f:
|
149 |
vqa = json.load(f)
|
150 |
+
if labeled:
|
151 |
+
for id_, d in enumerate(vqa):
|
152 |
+
img_features = id2features[d["img_id"]]
|
153 |
+
ids = [self.ans2label[x] for x in d["label"].keys()]
|
154 |
+
weights = list(d["label"].values())
|
155 |
+
yield id_, {
|
156 |
+
"question": d["sent"],
|
157 |
+
"question_type": d["question_type"],
|
158 |
+
"question_id": d["question_id"],
|
159 |
+
"image_id": d["img_id"],
|
160 |
+
"features": img_features["features"],
|
161 |
+
"normalized_boxes": img_features["normalized_boxes"],
|
162 |
+
"answer_type": d["answer_type"],
|
163 |
+
"label": {
|
164 |
+
"ids": ids,
|
165 |
+
"weights": weights,
|
166 |
+
},
|
167 |
+
}
|
168 |
+
else:
|
169 |
+
for id_, d in enumerate(vqa):
|
170 |
+
img_features = id2features[d["img_id"]]
|
171 |
+
yield id_, {
|
172 |
+
"question": d["sent"],
|
173 |
+
"question_type": "",
|
174 |
+
"question_id": d["question_id"],
|
175 |
+
"image_id": d["img_id"],
|
176 |
+
"features": img_features["features"],
|
177 |
+
"normalized_boxes": img_features["normalized_boxes"],
|
178 |
+
"answer_type": "",
|
179 |
+
"label": {
|
180 |
+
"ids": [],
|
181 |
+
"weights": [],
|
182 |
+
},
|
183 |
+
}
|