dataloader.py - what'a the output of this "labels"?
michaeldengxyz opened this issue · comments
Inside dataloader.py:
img_files = glob.glob(dataset_dir)
img_files = [f for f in img_files if f[-3:] in IM_EXTENSIONS]
with open(label_path, "rb") as f:
labels = pickle.load(f)
numofData = len(img_files)# endwiths(png,jpg ...)
data_idx = np.arange(numofData)
labels = pickle.load(f)
what'a the output of this array "labels"?
The labels are differed by datasets and how you define the anchor boxes.
I didn't make this repo to fit in a specific dataset. If you propose a dataset for an example, I would be able to answer your question.
Hello, I’d like to ask you something about training data and labels. What data do you use to train the model? Is the link available ?
Hello, I used datasets that I make on my own. This code is kind of template to utilize for various datasets. Thank you for asking!
The labels are differed by datasets and how you define the anchor boxes.
I didn't make this repo to fit in a specific dataset. If you propose a dataset for an example, I would be able to answer your question.
Hello, I have a dataset made by myself. the label is like (x, y, w, h), which means the left-top point, width and height of face in picture. I wonder how should I set the format of 'labels' in
with open(label_path, "rb") as f:
labels = pickle.load(f)
I would appreciate it if you give me a little example, Thank you !
@Jaren1907
Basically, BlazeFace uses Anchor box. Hence, the label form should follow Anchor scheme.
To make anchor boxes, I recommend you to refer RetinaNet repo (https://github.com/fizyr/keras-retinanet)
@minus31 I stil have no idea about the setting of labels, can you give me an example?
Hello, I have a dataset made by myself. the label is like (x, y, w, h), which means the left-top point, width and height of face in picture. I wonder how should I set the format of 'labels' in，Can you give details？thank you！