Open thuwzy opened 1 year ago
My result of hamburger, which is not as good as the one in the paper.
Would also be interested in this!
@thuwzy I found out that you can use the latent-tune mode (documented in the paper not the ReadMe) to receive better results. Simply set your nerf_type: latent_tune in the config.yaml and adjust your train_w/h and eval_w/h accordingly. You also have to fix the code according to this and beware that this will take up way more memory!
@JDihlmann Thank you for your kind response!
I tried to reproduce the result with default parameters. However, I got worse results than the reported results.