creotiv / hdrnet-pytorch

Unofficial PyTorch implementation of 'Deep Bilateral Learning for Real-Time Image Enhancement', SIGGRAPH 2017 https://groups.csail.mit.edu/graphics/hdrnet/
224 stars 43 forks source link

About the pre-trained model #28

Open Ssakura-go opened 2 years ago

Ssakura-go commented 2 years ago

Hello, dear friend,

After reading your paper [ Deep Bilateral Learning for Real-Time Image Enhancement ], I'd like to try to get the results on my own datasets. To verify whether I run on the right or not, I'd like to try on a pre-trained model. However, I couldn't find the pre-trained model.

I know that it's an unofficial implement, but if you are willing to, could you please share the weight what you get with me?

Thanks♪(・ω・)ノ

creotiv commented 2 years ago

1) it's not my paper 2) there is no pretrained models, as there still problems with network training (no one knows whys)

Kindly yours, Andrey Nikishaev

Areas ML/DS/CV/Soft Dev/BizDev/Growth Hacking/Customer Rel/IT LinkedIn http://ua.linkedin.com/in/creotiv GitHub http://github.com/creotiv Slideshare https://www.slideshare.net/anikishaev/ AnimalHelp https://patreon.com/uah Telegram @anikishaev Mobile +380632410666

On Mon, Aug 1, 2022 at 12:08 PM Ssakura-go @.***> wrote:

Hello, dear author,

After reading your paper [ Deep Bilateral Learning for Real-Time Image Enhancement ], I'd like to try to get the results on my own datasets. To verify whether I run on the right or not, I'd like to try on a pre-trained model. However, I couldn't find the pre-trained model.

I know that it's an unofficial implement, but if you are willing to, could you please share the weight what you get with me?

Thanks♪(・ω・)ノ

— Reply to this email directly, view it on GitHub https://github.com/creotiv/hdrnet-pytorch/issues/28, or unsubscribe https://github.com/notifications/unsubscribe-auth/AAB5CDNWVS4U5EJPM6RZNXLVW6HYHANCNFSM55GZJ7NQ . You are receiving this because you are subscribed to this thread.Message ID: @.***>