chainercv
chainercv copied to clipboard
chainer
ChainerCV: a Library for Deep Learning in Computer Vision
For example, the list `ade20k_semantic_segmentation_label_names` has two "tree" https://github.com/chainer/chainercv/blob/master/chainercv/datasets/ade20k/ade20k_utils.py#L27 https://github.com/chainer/chainercv/blob/master/chainercv/datasets/ade20k/ade20k_utils.py#L95 Some values are different from the original in https://github.com/CSAILVision/sceneparsing.
We now use tables for the documentation of dataset return values. We can use similar style for the return values of `predict`, which is often repetitive.
Segmented image can be displayed with vis_semantic_segmentation function. It can be saved as a result. What I suggest is a separate function that can return the segmented image directly. This...
https://github.com/chainer/chainercv/blob/9fe62c88fa034955b9ac4cfb0a6952b7c7354459/chainercv/visualizations/vis_point.py#L9 I think it's nice if it can take instance_colors in a similar manner to vis_box.
The function `mask_head_loss_pre` takes `gt_bboxes` as the 4th positional argument, but the docstring does not contain what this argument is. https://github.com/chainer/chainercv/blob/master/chainercv/links/model/fpn/mask_head.py#L171
Hi I have been trying to load my trained FPN weights and it shows this error, ```python raise KeyError("%s is not a file in the archive" % key) KeyError: 'extractor/base/res2/a/conv3/bn/beta...
This patch enables users to pass a `pathlib.Path` object to `read_image`, which is currently not possible due to the limitations `cv2.imread` ( `cv2.imread(pathlib.Path('hoge.png'))` throws a cryptic error )
For back compatibility, `read_image` 's `alpha` parameter should be 'blend_with_white'. Current `None` fails classifier train example, because imagenet has images with alpha.
Hi, I want to know that the value 0.0 of `neg_iou_thresh_lo` is intended or not. In the Fast R-CNN paper, it seems that the assignment for negative samples is decided...
Hi, Thank you for your great library. I'm training SSD model based on your example code (https://github.com/chainer/chainercv/blob/master/examples/ssd/train.py) . This example code requires at least one object annotated for each data...
