Issues
- 0
RGB 和BGR以及mean，std
#33 opened by Lyan-ing - 5
note: you need convert the original pretrained weights to d2 format by convert_to_d2.py
#11 opened by G-Apple1 - 0
- 2
My version of Detectron2
#30 opened by salehnia - 0
- 0
Instance Segmentation
#29 opened by Treeboy2762 - 1
训练Retinanet，提示This error indicates that your module has parameters that were not used in producing loss.
#28 opened by AutomanHan - 0
如何运行？
#27 opened by Wind010321 - 6
faster_rcnn训练ap为nan
#25 opened by hejinsome - 0
faster_rcnn训练ap为nan ，点击错了，希望重新打开 #25
#26 opened by hejinsome - 1
resnet related configuration in the yaml file
#24 opened by zhouzhouhhh - 1
Swintransformer结构问题
#23 opened by simonsmj - 10
version of torch
#20 opened by salehnia - 2
- 1
- 1
Pretrained model
#16 opened by VGrondin - 1
是否可以使用torch.jit.trace 成功呢
#12 opened by 13354236170 - 5
Sudden increment in loss values during training
#19 opened by purvang3 - 1
pred_masks
#18 opened by hrnavy - 4
GPU requirement
#17 opened by Jerrrrykun - 1
I want to check if the model is trained well so I do inference but in cfg.merge_from_file, MODEL.SWINT Non-existent error comes up again
#9 opened by leesangjoon1 - 5
- 0
Images random resized
#15 opened by mxin262 - 0
ModuleNotFoundError: No module named 'timm'
#14 opened by mxin262 - 0
About eval
#13 opened by Xuguozi - 1
About the train
#10 opened by G-Apple1 - 5
Faster R-CNN Pre-trained Model
#5 opened by Alchemistyui - 2
Detectron2 version
#6 opened by tandes - 2
- 3
RGB order and pixel mean/std?
#1 opened by tonysy