search

city-super / BungeeNeRF

[ECCV22] BungeeNeRF: Progressive Neural Radiance Field for Extreme Multi-scale Scene Rendering
https://city-super.github.io/citynerf
MIT License
565 stars 61 forks source link

Why does it have poor quality for my train and render #6

Closed hzhshok closed 2 years ago

hzhshok commented 2 years ago

Hello, thanks for you sharing the fantastic work. Now i download the example data and following the demo, but the the output 11 png images are not as the paper show, i don't know if my configuration parameter are not optimizated? the following is my configuration,

Could you help give me a suggestion?

Event the render quality is lower than the original images.

Hardware: GPU: 3080(24G)

Configuration: using default - 200000 iteraters.

Regards 011

Total 11 images.

kam1107 commented 2 years ago

See here: https://github.com/city-super/BungeeNeRF/issues/3#issuecomment-1200134284

hzhshok commented 2 years ago

@kam1107 Thanks lots for your quick response！

i will try that to get more level detailed renders.

Wei-Baldwin-Zeng commented 2 years ago

Do your render results getting better?

My render results look like yours, blurry around edges and loss details.

hzhshok commented 2 years ago

Yes, it is better the first one to my second step, and i just run two steps because the speed is very too slow and cost more machine performance, at machine free i could do a further training to check if it has good quality.

atztao commented 2 years ago

@hzhshok Hey, how long is the runtime ?