-
Notifications
You must be signed in to change notification settings - Fork 5
/
__environment_setup.txt
68 lines (43 loc) · 2.18 KB
/
__environment_setup.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
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
# Alternative running via Docker
- install docker
- and nvidia-docker (instructions from https://docs.nvidia.com/datacenter/cloud-native/container-toolkit/latest/install-guide.html seem to work)
- then download and run my image ...
# Install step 1: CUDA
(stick to the well known combo!)
(we used combination: cuda_10.0.130_410.48 + libcudnn7_7.4.2.24-1+cuda10.0_amd64.deb + tensorflow-gpu==1.14.0)
# Install step 2: Libraries
sudo apt-get install python3-opencv
sudo apt-get install libasound2-plugins
conda create -c conda-forge -n gan_explorer python=3.6 mamba
source activate gan_explorer
mamba install -c conda-forge requests==2.24.0 matplotlib==3.3.2 Pillow==8.0.1 mock==4.0.2 scikit-image==0.17.2
pip install opencv-python==4.4.0.46 oscpy==0.5.0
pip install tensorflow-gpu==1.14.0
(skipping) pip install pygame
(last step - download the dnnlib from stylegan2 repo)
\\\\\\\\\\\\\\\\\\\\\\\\\\\
# Runner codes:
[[[ GAN EXPLORER ]]]
##source /home/vitek/cuda-10.0-env
source activate gan_explorer
python demo.py -network models/sg_walk35k_network-snapshot-001841-good.pkl -architecture "StyleGAN2"
python demo.py -network models/AERIALS/pg_aerials512vectors1024px_snapshot-010200.pkl
# NICE DEMOS:
AERIALS
python demo.py -fullscreen full -network models/AERIALS/pg_aerials512vectors1024px_snapshot-010200.pkl
BUS
(15fps) python demo.py -fullscreen full -network models/BUS/pg_bus-snapshot-010300.pkl
python demo.py -fullscreen full -network models/BUS/sg_bus35k_network-snapshot-001882-good.pkl -architecture "StyleGAN2"
BOAT
[*](15fps) python demo.py -fullscreen full -network models/BOAT/pg_boat-snapshot-010300.pkl
(too fast~) python demo.py -fullscreen full -network models/BOAT/sg_boat35k_network-snapshot-001964-good.pkl -architecture "StyleGAN2"
CCTV DPP:
python demo.py -fullscreen full -network models/CCTV/pg_cctv_prague_dpp1-011000.pkl
[*]CCTV RUZYNE:
python demo.py -fullscreen full -network models/CCTV/pg_cctv_prague_Ruzyne_512_30k-009600.pkl
CCTV Vaclavak
python demo.py -fullscreen full -network models/CCTV/pg_cctv_prague_vaclavak_1024_30k-010300.pkl
(
TRAIN
python demo.py -fullscreen full -network models/TRAIN/___sg_train_side_network-snapshot-000982.pkl -architecture "StyleGAN2"
)