Chandler Timm Doloriel

Grayscale to Colored using CycleGAN

1

Can you check if my config is correct, I followed the instructions in the docs but just to be sure since this is my first time using GAN. The things...

How to replicate attention maps in object detection

5

Can you share the code on how to visualize attention maps in object detection like the one shown in your paper? 

How to use Multi-Head Attention in ViT

1

I need help to understand the multihead attention in [ViT](https://github.com/lucidrains/vit-pytorch/blob/main/vit_pytorch/vit.py). ``` class Attention(nn.Module): def __init__(self, dim, heads = 8, dim_head = 64, dropout = 0.): super().__init__() inner_dim = dim_head *...

PVT for Faster RCNN

2

Can I try your PVT and PVTv2 to Faster-RCNN? Will it also give a big improvement (if you tried simulating this)?

What could be the best learning rate if I have 1 GPU, and I set the samples_per_gpu=1 and workers_per_gpu=1?

Visualizing Attention Map

1

Is there a way to visualize the attention maps? Have you tried it? In vision transformers, we can easily visualize attention maps but for your paper can we do it?

The DOTA dataset annotations are specified in corner positions (x1, y1, x2, y2, x3, y3, x4, y4) but why did you define your bounding box in center+height+width (x,y,w,h) position? Does...

Distributed Multi-GPU Training did not decrease training time

1

When I tried distributed training for `2 RTX A100 GPU's` with `batch size of 4 images per GPU`, the training time did not decrease. When I change `batch size to...

@jbwang1997 How can I edit the below config to load `gt_masks` for DOTA dataset? ```python train_pipeline = [ dict(type='LoadImageFromFile'), dict(type='LoadOBBAnnotations', with_bbox=True, with_label=True, obb_as_mask=True), dict(type='LoadDOTASpecialInfo'), dict(type='Resize', img_scale=(1024, 1024), keep_ratio=True), dict(type='OBBRandomFlip', h_flip_ratio=0.5,...

It's been a while since I read your great paper. This is what I remembered: If I'm right, this is a direct improvement of `RoI transformer` (and its variant `ReDet`)...