Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

About the mclight part #9

Open
ODRLver opened this issue Mar 28, 2024 · 2 comments
Open

About the mclight part #9

ODRLver opened this issue Mar 28, 2024 · 2 comments

Comments

@ODRLver
Copy link

ODRLver commented Mar 28, 2024

Hi! Thanks for the amazing work!
I am really interested in your multi-camera images to HDR skydome part. I tried to run the infer.py in the mc_to_sky/tools files, and I copied the config info from the Pretrain zip you provided. The infer part worked well after I ignored all the request to import a unet module from the sub_model file. I run some test based on the waymo segment: "segment-11379226583756500423_6230_810_6250_810_with_camera_labels", and I imported the exr documents end without '_sky' into the blender, the shadow direction doesn't seem to be right, am I setting the config wrong? And when I change the config to infer from only a single image, the rendered image from blender seems to lack the sunshine. I couldn't help but noticed the line written on top of the infer.py, it gives the command to infer the hdr from a single image, but I can't fine the config '/home/yfl/workspace/LDR_to_HDR/mc_to_sky/logs/pred_hdr_pano_from_single_1012_195149/config.yaml' maybe you could post it?
Thanks in advance!

@yifanlu0227
Copy link
Owner

Hi! We predict the skydome HDRI for each frame and select the best one. The scene 113792 is a little difficult and we find the 033.exr performs best.

You can leave me an email address and I would send you the code snippet.

@ODRLver

This comment was marked as off-topic.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

No branches or pull requests

2 participants