Satoshi Tsutsui

Thanks for the good implementation! I was a bit confused about the description so I'd like to comment. > Then in each epoch, the loader will sample the entire dataset...

pixel_aspect_ratio is always one because of `pixel_aspect_ratio=colmap_camera.fx / colmap_camera.fx` - see https://github.com/google/nerfies/blob/5b181d8f50fa961139f86aa31f2e05c9190aaec4/notebooks/Nerfies_Capture_Processing.ipynb ``` def convert_colmap_camera(colmap_camera, colmap_image): """Converts a pycolmap `image` to an SFM camera.""" camera_rotation = colmap_image.R() camera_position = -(colmap_image.t...

[suggestion] remove chainermn dependency from FPN sample.

7

I mentioned here https://github.com/chainer/chainercv/issues/735#issuecomment-479616802_ before and currently FPN detector is depending on the chainermn. Unfortunately chainermn is not easy to install for those (including me) who are not familiar with...

I can install chainer and cupy via conda command but can't do for chainercv. How about make it installable via conda? I mean, This works `conda install chainer` `conda install...

vis_point cannot accept user defined colors for each point.

1

https://github.com/chainer/chainercv/blob/9fe62c88fa034955b9ac4cfb0a6952b7c7354459/chainercv/visualizations/vis_point.py#L9 I think it's nice if it can take instance_colors in a similar manner to vis_box.

implementation for the segmentation?

4

Hi, I could not find the implementation for the segmentation and its evaluation.... Where is it? Best, Satoshi

Thanks for the nice implementation! It would be nice if you make it comparable with accimage ( https://github.com/pytorch/accimage ).... Currently it assume numpy or Pillow image.

why the model use only first three channels of the last layer output ?

6

https://github.com/huggingface/pytorch-pretrained-BigGAN/blob/1e18aed2dff75db51428f13b940c38b923eb4a3d/pytorch_pretrained_biggan/model.py#L245-L246 I'm trying to understand the model by reading code. I noticed that `conv_to_rgb` has actually 128 channels but only first three are used for the final RGB image. Why...

Crawling script

3

Hi, Yanghua, Can you please open sources the script you used for crawling from www.getchu.com ? In this way, you can avoid copyright issue. Best Regards, Satoshi

Make it work with cpu only

1

Nice work! It's very light (only 4MB of weights) tool to detect hands! I don't have cuda on my laptop and so wanted to run without cuda. I made it,...