Sanghoon Hong

Hi @XiongweiWu, I'm sorry to say that it's complicated to locate the old PVA-lite model on ImageNet. (You can still download the PVA-lite model for detection) However, we have a...

Hi @xiaoxiongli, I guess maybe it's due to limited computational costs. Could you share some quantitative comparison results?

Hi @beihangzxm123 , I'm not sure how big your dataset is, but if it's not that big I recommend you do ... - start from the up-to-date PVANet detection model...

Since our target was to develop a detector for few selected objects, we haven't evaluated our models on the MS COCO (80 class) dataset.

Hi @twmht I've done it when I worked for Intel, but now I work for KakaoBrain. I'm thinking of reproducing this result with PyTorch, but I'm now sure whether I'm...

Hi @mingde510 Did you set the cfg argument?

Hi @cloudyszz Have you added an argument '--cfg models/pvanet/cfgs/submit_0716.yml' or '--cfg models/pvanet/cfgs/submit_1019.yml' ?

Hi @kaishijeng I may be able to share a new lite model with prototxts in a couple of weeks.

Sorry for a late comment. I'm attending a conference right now, I'll check the issue when I get back to my home. However, I guess it may be a training...

Hi @catsdogone We've added C_ReLU mainly for test efficiency. I'm not 100% sure whether it can accelerate network training. And as @songjmcn mentioned, it'll depend on implementations