IDEA-Research/DINO
[ICLR 2023] Official implementation of the paper "DINO: DETR with Improved DeNoising Anchor Boxes for End-to-End Object Detection"
PythonApache-2.0
Issues
- 1
Question about the FPS of DINO-SwinL model?
#250 opened by WUSHUANGPPP - 0
Will O365 pretrained weights be released?
#261 opened by h-bo - 0
visualizing bounding box
#260 opened by 01fe21bec239 - 0
cuda not available
#259 opened by 01fe21bec239 - 0
Msda module not found
#258 opened by 01fe21bec239 - 1
Error while training
#257 opened by 01fe21bec239 - 1
How to restart fine-tuning where I left off
#256 opened by 5starkarma - 4
train error
#251 opened by kkkkkk123-ops - 2
Inference scale on DINO model
#242 opened by truong2710-cyber - 2
- 0
data_aug_scale_overlap
#255 opened by Bill-WangJiLong - 0
- 0
- 0
size mismatch.My num_classes is 4, and I want to use dino's checkpoint, which causes an error.
#239 opened by xqq777 - 9
The AP value is low
#208 opened by 1106X - 0
num_classes
#249 opened by yuzhixiaobo - 0
swinv2 backbone Causing map to be 0 during training and inference process
#248 opened by everytimeis80 - 0
COCO test-dev
#247 opened by yaw91 - 0
关于two-stage topk_proposals的问题
#245 opened by nhw649 - 0
MultiScaleDe
#244 opened by YellowOrz - 0
- 0
reproducibility
#241 opened by LOVE-YOURSELF-1 - 0
- 0
How is top-k calculated?
#238 opened by yangtle - 5
- 1
MultiScaleDeformableAttention complied but ImportError undefined symbol: THPVariableClass
#237 opened by RiceZ - 0
Name of hyper-parameters λ1 and λ2
#236 opened by adsbansal - 0
flops of DINO_r50
#235 opened by mc-lgt - 2
fine tuning on custom dataset
#234 opened by zyx20180209 - 0
- 0
evaluation with different data_loader_val batch size leads to different results
#231 opened by cdluminate - 0
denoising groups
#230 opened by NeKoooo233 - 1
how to inference a video ?
#228 opened by IronmanVsThanos - 0
how to picke a good epoch to drop lr
#229 opened by onepeachbiubiubiu - 1
mAP = 48.6
#223 opened by wulele2 - 1
- 1
- 0
- 0
Question about the number of PatchMerging when swin-transformer is used as a backbone.
#224 opened by Kim-yhao - 0
Prediction of 4K images using default parameters and weights from COCO dataset
#222 opened by pascual-tejero - 0
Is DINO compatible with 3D datasets?
#219 opened by MadhuSaran26 - 0
Never predicts 'no-object'
#218 opened by JeremyRss - 0
Add pyproject.toml for python packaging
#217 opened by yuho-jeong - 0
Weights/model .pth file for Swin-L + DINO pretrained on Objects365 and fine-tuned on COCO.
#216 opened by shivasv97 - 0
fail run train on device 1
#215 opened by xbkaishui - 2
- 1
大佬,您好! 性能最优
#209 opened by sdreamforchen - 1
Increasing batch size does not increase inference speed per image, and negatively impacts mAP
#212 opened by InternetXplorer - 0
Unstable training with dense scenes
#213 opened by half-truism - 0
Object365 pretrain
#211 opened by zhenliulv