QueryInst

QueryInst copied to clipboard

i use mmcv 1.3.3 and 1.4，there will be an assertion: ca_forward miss in module _ext so, what's your version of mmcv test.py used? or maybe there are some other reasons?

Judy-Liang

您好，请问代码里是不是没使用预训练模型？

1

seekerwmy

Why use 'batch-nms' in an end2end model?

niecongchong

mAP = 0 all the time

7

In the training log, the mAP is 0 all the time bbox_mAP: 0.0000, bbox_mAP_50: 0.0000, bbox_mAP_75: 0.0000, bbox_mAP_s: -1.0000, bbox_mAP_m: -1.0000, bbox_mAP_l: 0.0000, bbox_mAP_copypaste: 0.000 0.000 0.000 -1.000 -1.000 0.000,...

Kangzf1996

cityscapes config

10

你好，我是一个研究生，最近在复现你的算法，可以具体告诉我一下cityscapes下的配置文件细节吗，非常感谢！ doge

zhushilinn

fix: correct the problem when fcn_mask_head takes in invalid bboxes with negative coordinates

1

#41 Main edit was at the original line 359 - 362

eric-kwok-nt

The error that I have encountered when inferencing using the swin transformer on CPU was `ERROR - upper bound and larger bound inconsistent with step sign` Such error disappears when...

eric-kwok-nt

how to inference the pretrained model on my test data without GT

1

Hi, guys. How to inference my own data without GT?

linchunmian

Hi, thanks for the great work! I tried to use QueryInst_Swin_L_300_queries (single scale testing) coco with mmdetection. I download the linked config file `queryinst_swin_large_patch4_window7_fpn_300_proposals_crop_mstrain_400-1200_50e_coco.py` but unfortunately the base config it...

bananajoe182

gpu

2

请问一张v100能训练吗

roar-1128