neuralangelo记录

前言

提心吊胆地跑nerualangelo，不仅是因为github上的tnt出问题的人太多了，还担心motivation验证不出来吧。

training

config

# 修改数据集目录root:datasets/tanks_and_temples/Truck num_images:251# The number of training images.

mesh extraction

GROUP=tnt NAME=truck CHECKPOINT=logs/${GROUP}/${NAME}/epoch_01992_iteration_000500000_checkpoint.pt OUTPUT_MESH=logs/${GROUP}/${NAME}/${NAME}.ply CONFIG=logs/${GROUP}/${NAME}/config.yaml RESOLUTION=2048 BLOCK_RES=128 GPUS=1 # use >1 for multi-GPU mesh extraction torchrun --nproc_per_node=${GPUS} projects/neuralangelo/scripts/extract_mesh.py \ --config=${CONFIG} \ --checkpoint=${CHECKPOINT} \ --output_file=${OUTPUT_MESH} \ --resolution=${RESOLUTION} \ --block_res=${BLOCK_RES}

GROUP=nerf-syntheticNAME=nerf_legoRESOLUTION=1024BLOCK_RES=128CONFIG=logs/${GROUP}/${NAME}/config.yamlGPUS=1CHECKPOINT=logs/${GROUP}/${NAME}/epoch_00250_iteration_000100000_checkpoint.ptOUTPUT_MESH=logs/${GROUP}/${NAME}/lego1w.ply torchrun--nproc_per_node=${GPUS}projects/neuralangelo/scripts/extract_mesh.py--config=${CONFIG}--checkpoint=${CHECKPOINT}--output_file=${OUTPUT_MESH}--resolution=${RESOLUTION}--block_res=${BLOCK_RES}

from torch_scatterimportsegment_coo pipinstalltorch-scatter-fhttps://data.pyg.org/whl/torch-2.11.0+cu128.html