changed chunk generation
Browse files- ref_seg_ger.py +12 -7
ref_seg_ger.py
CHANGED
@@ -115,7 +115,7 @@ def load_image(image_path, size=None):
|
|
115 |
class RefSeg(datasets.GeneratorBasedBuilder):
|
116 |
"""TODO: Short description of my dataset."""
|
117 |
|
118 |
-
CHUNK_SIZE =
|
119 |
VERSION = datasets.Version("1.0.0")
|
120 |
|
121 |
# This is an example of a dataset with multiple configurations.
|
@@ -255,12 +255,17 @@ class RefSeg(datasets.GeneratorBasedBuilder):
|
|
255 |
if input != '':
|
256 |
clean_input_ids.append(input)
|
257 |
clean_labels.append(labels[i])
|
258 |
-
|
259 |
-
|
|
|
|
|
|
|
|
|
|
|
260 |
#split_bboxes = bboxes[index:index + self.CHUNK_SIZE]
|
261 |
# split_rgbs = rgbs[index:index + self.CHUNK_SIZE]
|
262 |
# split_fonts = fonts[index:index + self.CHUNK_SIZE]
|
263 |
-
split_labels = clean_labels[index:max(len(clean_input_ids), index + self.CHUNK_SIZE)]
|
264 |
#split_labels_post = [item for sublist in split_labels for item in sublist]
|
265 |
# if(len(split_ids) != len(split_labels)):
|
266 |
# print(f)
|
@@ -278,13 +283,13 @@ class RefSeg(datasets.GeneratorBasedBuilder):
|
|
278 |
|
279 |
yield key, {
|
280 |
#"id": f"{os.path.basename(f)}_{chunk_id}",
|
281 |
-
'input_ids':
|
282 |
-
'attention_mask': [1] * len(
|
283 |
#"bbox": split_bboxes,
|
284 |
# "RGBs": split_rgbs,
|
285 |
# "fonts": split_fonts,
|
286 |
#"image": image,
|
287 |
#"original_image": original_image,
|
288 |
-
"labels":
|
289 |
}
|
290 |
key += 1
|
|
|
115 |
class RefSeg(datasets.GeneratorBasedBuilder):
|
116 |
"""TODO: Short description of my dataset."""
|
117 |
|
118 |
+
CHUNK_SIZE = 256
|
119 |
VERSION = datasets.Version("1.0.0")
|
120 |
|
121 |
# This is an example of a dataset with multiple configurations.
|
|
|
255 |
if input != '':
|
256 |
clean_input_ids.append(input)
|
257 |
clean_labels.append(labels[i])
|
258 |
+
n_chunks = int(len(clean_input_ids)/self.CHUNK_SIZE) if len(clean_input_ids)%self.CHUNK_SIZE == 0 \
|
259 |
+
else int(len(clean_input_ids)/self.CHUNK_SIZE) + 1
|
260 |
+
split_ids = np.array_split(clean_input_ids, n_chunks)
|
261 |
+
split_labels = np.array_split(clean_labels, n_chunks)
|
262 |
+
for chunk_ids, chunk_labels in zip(split_ids, split_labels):
|
263 |
+
#for chunk_id, index in enumerate(range(0, len(clean_input_ids), self.CHUNK_SIZE)):
|
264 |
+
#split_ids = clean_input_ids[index:max(len(clean_input_ids), index + self.CHUNK_SIZE)]
|
265 |
#split_bboxes = bboxes[index:index + self.CHUNK_SIZE]
|
266 |
# split_rgbs = rgbs[index:index + self.CHUNK_SIZE]
|
267 |
# split_fonts = fonts[index:index + self.CHUNK_SIZE]
|
268 |
+
#split_labels = clean_labels[index:max(len(clean_input_ids), index + self.CHUNK_SIZE)]
|
269 |
#split_labels_post = [item for sublist in split_labels for item in sublist]
|
270 |
# if(len(split_ids) != len(split_labels)):
|
271 |
# print(f)
|
|
|
283 |
|
284 |
yield key, {
|
285 |
#"id": f"{os.path.basename(f)}_{chunk_id}",
|
286 |
+
'input_ids': chunk_ids,
|
287 |
+
'attention_mask': [1] * len(chunk_ids),
|
288 |
#"bbox": split_bboxes,
|
289 |
# "RGBs": split_rgbs,
|
290 |
# "fonts": split_fonts,
|
291 |
#"image": image,
|
292 |
#"original_image": original_image,
|
293 |
+
"labels": chunk_labels
|
294 |
}
|
295 |
key += 1
|