Ai作图工业化流程工具_clipvision model

背景：

今天给大家介绍一款新的工具，用过UE5、Blender的朋友对蓝图应该不陌生。蓝图绝对是UE5在游戏、虚拟制作、虚拟摄影爆火的大杀器。这工具把游戏制作直接变成了拖拖拉拽拽的工作，让艺术家、地形师、灯光师、渲染人员和底层技术人员可以一起协同作业。把游戏这个看起来toy的行业变成了可以写协作流程化的工作。

吹牛吹了这么久，此处开始切入正题。今天给大家介绍一套基于stablediffusion作图的蓝图工具。这套工具和stablediffusion webui的差异在于 stablediffusion wenui更适合一个人作业，工具更扁平更适合懂代码的朋友。对于真正把作图当成一个流水线，需要多到工序协作的小团队作业不算太优化。

ComfyUI一款为AI作图流水线化而生的产品，还处在产品打磨阶段，但已经可以看到它的牛逼。

工具介绍

长什么样

通过拖拉拽增加节点处理流程，看起来挺酷炫。

如何搭建

初始化环境

!git clone https://github.com/comfyanonymous/ComfyUI
%cd ComfyUI
!pip install xformers -r requirements.txt

下载软件

# Checkpoints
 
# SD1.5
!wget -c https://huggingface.co/runwayml/stable-diffusion-v1-5/resolve/main/v1-5-pruned-emaonly.ckpt -P ./models/checkpoints/
 
# SD2
#!wget -c https://huggingface.co/stabilityai/stable-diffusion-2-1-base/resolve/main/v2-1_512-ema-pruned.safetensors -P ./models/checkpoints/
#!wget -c https://huggingface.co/stabilityai/stable-diffusion-2-1/resolve/main/v2-1_768-ema-pruned.safetensors -P ./models/checkpoints/
 
# Some SD1.5 anime style
#!wget -c https://huggingface.co/WarriorMama777/OrangeMixs/resolve/main/Models/AbyssOrangeMix2/AbyssOrangeMix2_hard.safetensors -P ./models/checkpoints/
#!wget -c https://huggingface.co/WarriorMama777/OrangeMixs/resolve/main/Models/AbyssOrangeMix3/AOM3A1_orangemixs.safetensors -P ./models/checkpoints/
#!wget -c https://huggingface.co/WarriorMama777/OrangeMixs/resolve/main/Models/AbyssOrangeMix3/AOM3A3_orangemixs.safetensors -P ./models/checkpoints/
#!wget -c https://huggingface.co/Linaqruf/anything-v3.0/resolve/main/anything-v3-fp16-pruned.safetensors -P ./models/checkpoints/
 
# Waifu Diffusion 1.5 (anime style SD2.x 768-v)
#!wget -c https://huggingface.co/waifu-diffusion/wd-1-5-beta2/resolve/main/checkpoints/wd-1-5-beta2-fp16.safetensors -P ./models/checkpoints/
 
 
# VAE
!wget -c https://huggingface.co/stabilityai/sd-vae-ft-mse-original/resolve/main/vae-ft-mse-840000-ema-pruned.safetensors -P ./models/vae/
#!wget -c https://huggingface.co/WarriorMama777/OrangeMixs/resolve/main/VAEs/orangemix.vae.pt -P ./models/vae/
#!wget -c https://huggingface.co/hakurei/waifu-diffusion-v1-4/resolve/main/vae/kl-f8-anime2.ckpt -P ./models/vae/
 
 
# Loras
#!wget -c https://civitai.com/api/download/models/10350 -O ./models/loras/theovercomer8sContrastFix_sd21768.safetensors #theovercomer8sContrastFix SD2.x 768-v
#!wget -c https://civitai.com/api/download/models/10638 -O ./models/loras/theovercomer8sContrastFix_sd15.safetensors #theovercomer8sContrastFix SD1.x
 
 
# T2I-Adapter
#!wget -c https://huggingface.co/TencentARC/T2I-Adapter/resolve/main/models/t2iadapter_depth_sd14v1.pth -P ./models/t2i_adapter/
#!wget -c https://huggingface.co/TencentARC/T2I-Adapter/resolve/main/models/t2iadapter_seg_sd14v1.pth -P ./models/t2i_adapter/
#!wget -c https://huggingface.co/TencentARC/T2I-Adapter/resolve/main/models/t2iadapter_sketch_sd14v1.pth -P ./models/t2i_adapter/
#!wget -c https://huggingface.co/TencentARC/T2I-Adapter/resolve/main/models/t2iadapter_keypose_sd14v1.pth -P ./models/t2i_adapter/
#!wget -c https://huggingface.co/TencentARC/T2I-Adapter/resolve/main/models/t2iadapter_openpose_sd14v1.pth -P ./models/t2i_adapter/
#!wget -c https://huggingface.co/TencentARC/T2I-Adapter/resolve/main/models/t2iadapter_color_sd14v1.pth -P ./models/t2i_adapter/
#!wget -c https://huggingface.co/TencentARC/T2I-Adapter/resolve/main/models/t2iadapter_canny_sd14v1.pth -P ./models/t2i_adapter/
 
# T2I Styles Model
#!wget -c https://huggingface.co/TencentARC/T2I-Adapter/resolve/main/models/t2iadapter_style_sd14v1.pth -P ./models/style_models/
 
# CLIPVision model (needed for styles model)
#!wget -c https://huggingface.co/openai/clip-vit-large-patch14/resolve/main/pytorch_model.bin -O ./models/clip_vision/clip_vit14.bin
 
 
# ControlNet
#!wget -c https://huggingface.co/webui/ControlNet-modules-safetensors/resolve/main/control_depth-fp16.safetensors -P ./models/controlnet/
#!wget -c https://huggingface.co/webui/ControlNet-modules-safetensors/resolve/main/control_scribble-fp16.safetensors -P ./models/controlnet/
#!wget -c https://huggingface.co/webui/ControlNet-modules-safetensors/resolve/main/control_openpose-fp16.safetensors -P ./models/controlnet/
 
 
# Controlnet Preprocessor nodes by Fannovel16
#!cd custom_nodes && git clone https://github.com/Fannovel16/comfy_controlnet_preprocessors; cd comfy_controlnet_preprocessors && python install.py
 
# ESRGAN upscale model
#!wget -c https://huggingface.co/sberbank-ai/Real-ESRGAN/resolve/main/RealESRGAN_x2.pth -P ./models/upscale_models/
#!wget -c https://huggingface.co/sberbank-ai/Real-ESRGAN/resolve/main/RealESRGAN_x4.pth -P ./models/upscale_models/

运行可视化界面

!npm install -g localtunnel
 
import subprocess
import threading
import time
import socket
def iframe_thread(port):
  while True:
      time.sleep(0.5)
      sock = socket.socket(socket.AF_INET, socket.SOCK_STREAM)
      result = sock.connect_ex(('127.0.0.1', port))
      if result == 0:
        break
      sock.close()
  print("\nComfyUI finished loading, trying to launch localtunnel (if it gets stuck here localtunnel is having issues)")
  p = subprocess.Popen(["lt", "--port", "{}".format(port)], stdout=subprocess.PIPE)
  for line in p.stdout:
    print(line.decode(), end='')
 
 
threading.Thread(target=iframe_thread, daemon=True, args=(8188,)).start()
 
!python main.py --dont-print-server

iframe打开方式

import threading
import time
import socket
def iframe_thread(port):
  while True:
      time.sleep(0.5)
      sock = socket.socket(socket.AF_INET, socket.SOCK_STREAM)
      result = sock.connect_ex(('127.0.0.1', port))
      if result == 0:
        break
      sock.close()
  from google.colab import output
  output.serve_kernel_port_as_iframe(port, height=1024)
  print("to open it in a window you can open this link here:")
  output.serve_kernel_port_as_window(port)
 
 
threading.Thread(target=iframe_thread, daemon=True, args=(8188,)).start()
 
 
!python main.py --dont-print-server

小结：

1.Ai作图已成态势，适合普通人的作图流水线

2.稳定的AI作图才能产生工业价值，而不是toy

3.蓝图是种很好的艺术家、技术家...不同工种协同的工具