forked from VITA-Group/GNT
-
Notifications
You must be signed in to change notification settings - Fork 0
/
record_gnt.txt
51 lines (38 loc) · 1.73 KB
/
record_gnt.txt
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
1. GNT:
GNT论文代码:https://github.com/VITA-Group/GNT
1.1 激活环境
打开anaconda图形界面: anaconda-navigator
查看已配置环境:conda env list
激活环境:source activate gnt
关闭环境:conda deactivate
1.2 训练测试模型
训练模型:
# single scene
python3 train.py --config configs/gnt_<scene-name>.txt --train_scenes <scene> --eval_scenes <scene>
eg:python3 train.py --config configs/gnt_llff.txt --train_scenes orchids --eval_scenes orchids
# cross scene
python3 train.py --config configs/gnt_full.txt
output:./GNT/out/gnt_<scene-name>/model_XXXXXX.pth(model)
测试模型:
# single scene
python3 eval.py --config configs/gnt_<scene-name>.txt --eval_scenes <scene> --expname gnt_<scene-name> --chunk_size 500 --run_val --N_samples 192
# cross scene
python3 eval.py --config configs/gnt_full.txt --expname gnt_full --chunk_size 500 --run_val --N_samples 192
渲染:python3 render.py --config configs/gnt_llff.txt --eval_dataset llff_render --eval_scenes <scene> --expname gnt_<scene-name> --chunk_size 500 --N_samples 192
output:120张图片train_%3d_coarse.png
拼合图片成为视频:ffmpeg -r 40 -f image2 -i train_%3d_coarse.png output1.mp4
1.3 输入输出
数据集构建:
├──data/
├──ibrnet_collected_1/ (*)
├──ibrnet_collected_2/ (*)
├──real_iconic_noface/
├──spaces_dataset/
├──RealEstate10K-subset/ (no)
├──google_scanned_objects/
├──nerf_synthetic/ (*)
├──nerf_llff_data/ (*)
训练模型存放:./out/{gnt_<scene-name>
1.4参数设置
configs:./configs/gnt_<scene-name>.txt
显卡显存不足可以减小参数:N_rand