Ways to Install this with Conda and Linux Server

[longlifedahan]

1、install conda
2、install git
3、create virtual environment
4、download pytorch=1.12.1 and cuda=11.6
5、download [stable-diffusion-webui]、[stable-diffusion]；[CLIP]、[k-diffusion]、[taming-transformers]、[DeepDanbooru]
6、unpack [stable-diffusion-webui](or use git clone instead)
7、create dir named "repositories"
8、unpack [stable-diffusion][CLIP][k-diffusion][taming-transformers][DeepDanbooru] in repositories
9、install all requirement of the upper 6 packages, but considering delete the version of pytorch and cuda in them(some are too old and may cause conflict)
10、put models /stable-diffusion-webui/models/Stable-diffusion/
11、goto /stable-diffusion-webui, activate virtual enviroment,and python webui.py**(it is really confusing that even if I downloaded pytorch and cuda via conda, the program downloaded again)**
12、ssh to the Linux Server Port 7860, then access it using windows
(p.s.)
13、using xformers will speed up picturing, to do this,you must:

• conda install xformers -c xformers/label/dev
• goto /miniconda3/pkgs/ and find xformers-0.0.14.dev315+git.e23b369-py310_cu11.3_pyt1.12.1.tar.bz2
• unpack it,and place the dir xformers in /stable-diffusion-webui
• use --xformers