Datasets:
Tasks:
Visual Question Answering
Formats:
parquet
Languages:
English
Size:
10K - 100K
ArXiv:
Tags:
medical
License:
Commit
·
33455bf
1
Parent(s):
a97487c
Upload processing.py
Browse files- scripts/processing.py +6 -1
scripts/processing.py
CHANGED
|
@@ -2,6 +2,7 @@
|
|
| 2 |
creates an "imagefolder" dataset and pushes it to the Hugging Face Hub.
|
| 3 |
"""
|
| 4 |
|
|
|
|
| 5 |
import os
|
| 6 |
import shutil
|
| 7 |
import pickle
|
|
@@ -17,6 +18,11 @@ for split in ["train", "val", "test"]:
|
|
| 17 |
|
| 18 |
# drop the duplicate image-question-answer triplets
|
| 19 |
data = data.drop_duplicates(ignore_index=True)
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 20 |
|
| 21 |
# copy the images using unique file names
|
| 22 |
data.insert(0, "file_name", "")
|
|
@@ -32,4 +38,3 @@ for split in ["train", "val", "test"]:
|
|
| 32 |
# push the dataset to the hub
|
| 33 |
dataset = datasets.load_dataset("imagefolder", data_dir="data/")
|
| 34 |
dataset.push_to_hub("flaviagiammarino/path-vqa")
|
| 35 |
-
|
|
|
|
| 2 |
creates an "imagefolder" dataset and pushes it to the Hugging Face Hub.
|
| 3 |
"""
|
| 4 |
|
| 5 |
+
import re
|
| 6 |
import os
|
| 7 |
import shutil
|
| 8 |
import pickle
|
|
|
|
| 18 |
|
| 19 |
# drop the duplicate image-question-answer triplets
|
| 20 |
data = data.drop_duplicates(ignore_index=True)
|
| 21 |
+
|
| 22 |
+
# perform some basic data cleaning/normalization
|
| 23 |
+
f = lambda x: re.sub(' +', ' ', str(x).lower()).replace(" ?", "?").strip()
|
| 24 |
+
data["question"] = data["question"].apply(f)
|
| 25 |
+
data["answer"] = data["answer"].apply(f)
|
| 26 |
|
| 27 |
# copy the images using unique file names
|
| 28 |
data.insert(0, "file_name", "")
|
|
|
|
| 38 |
# push the dataset to the hub
|
| 39 |
dataset = datasets.load_dataset("imagefolder", data_dir="data/")
|
| 40 |
dataset.push_to_hub("flaviagiammarino/path-vqa")
|
|
|