From 5511f5f4edf45a163a19ce2ff41e8b5fe2b9771e Mon Sep 17 00:00:00 2001 From: jMysliwiec <33121390+jMysliwiec@users.noreply.github.com> Date: Thu, 8 Jun 2023 17:22:46 +0200 Subject: [PATCH] Quick fix for eval script failing for depth-nerfacto (and other methods with a non-standard dataset) (#2047) dirty fix for the eval_utils script that fails because generics don't work with yaml configs Co-authored-by: Jakub Mysliwiec --- nerfstudio/utils/eval_utils.py | 3 +++ 1 file changed, 3 insertions(+) diff --git a/nerfstudio/utils/eval_utils.py b/nerfstudio/utils/eval_utils.py index 4d531ba45..acd8bee18 100644 --- a/nerfstudio/utils/eval_utils.py +++ b/nerfstudio/utils/eval_utils.py @@ -25,6 +25,7 @@ from typing import Literal, Optional, Tuple import torch import yaml +from nerfstudio.configs.method_configs import all_methods from nerfstudio.data.datamanagers.base_datamanager import VanillaDataManagerConfig from nerfstudio.engine.trainer import TrainerConfig from nerfstudio.pipelines.base_pipeline import Pipeline @@ -87,6 +88,8 @@ def eval_setup( config = yaml.load(config_path.read_text(), Loader=yaml.Loader) assert isinstance(config, TrainerConfig) + config.pipeline.datamanager._target = all_methods[config.method_name].pipeline.datamanager._target + if eval_num_rays_per_chunk: config.pipeline.model.eval_num_rays_per_chunk = eval_num_rays_per_chunk