File size: 5,143 Bytes
d0939b5 9032865 d0939b5 9032865 d0939b5 9032865 d0939b5 9032865 d0939b5 9032865 d0939b5 |
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 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 |
import os
from PIL import Image
import torchvision.transforms.functional as f
from utils import load_face_generator
from omegaconf import OmegaConf
import random
import sys
def generate_face_image(
anything_facemaker,
class_concept,
face_img_pil=None,
controlnet_conditioning_scale=1.0,
strength=0.95,
):
face_img_pil = f.center_crop(
f.resize(face_img_pil, 512), 512).convert('RGB')
prompt = anything_facemaker.prompt_template.format(class_concept)
# # There are four ways to generate a image by now.
# pure_generate = anything_facemaker.generate(prompt=prompt, image=face_img_pil, do_inversion=False)
# inversion = anything_facemaker.generate(prompt=prompt, image=face_img_pil, strength=strength, do_inversion=True)
if controlnet_conditioning_scale == None:
init_face_pil = anything_facemaker.generate(prompt=prompt)
return init_face_pil
if strength is None:
pure_control = anything_facemaker.face_control_generate(prompt=prompt, face_img_pil=face_img_pil, do_inversion=False,
controlnet_conditioning_scale=controlnet_conditioning_scale)
init_face_pil = pure_control
else:
control_inversion = anything_facemaker.face_control_generate(prompt=prompt, face_img_pil=face_img_pil, do_inversion=True,
strength=strength,
controlnet_conditioning_scale=controlnet_conditioning_scale)
init_face_pil = control_inversion
return init_face_pil
def experiment(anything_facemaker, concepts_path, face_img_path, output_dir,
controlnet_conditioning_scale=1., strength=0.95):
os.makedirs(output_dir, exist_ok=True)
face_img_pil = Image.open(face_img_path)
face_img_pil = f.center_crop(
f.resize(face_img_pil, 512), 512).convert('RGB')
with open(concepts_path) as fr:
concepts = fr.read().split('\n')
concepts = [concept for concept in concepts if len(concept)!=0]
random.shuffle(concepts)
for concept in concepts[:4]:
save_path = os.path.join(output_dir, f'{concept}.png')
if os.path.exists(save_path):
continue
init_face_pil = generate_face_image(
anything_facemaker,
class_concept=concept,
face_img_pil=face_img_pil,
controlnet_conditioning_scale=controlnet_conditioning_scale,
strength=strength,
)
save_path = os.path.join(output_dir, f'{concept}.png')
init_face_pil.save(save_path)
if __name__=='__main__':
# run this in repo path:
# PYTHONPATH=.:$PYTHONPATH python experiments/experiment.py
model_config_path = 'resources/models.yaml'
# model_config_path = 'resources/models_personality.yaml'
model_config = OmegaConf.load(model_config_path)['models']
gameicon_config = model_config['GameIconInstitute_mode']
# face_img_path = 'resources/images/faces/0.jpg'
face_img_dir='resources/images/faces'
faces = os.listdir(face_img_dir)
controlnet_conditioning_scale=1.
strength=0.95
for model, model_info in model_config.items():
anything_facemaker = load_face_generator(
model_dir=model_info['model_dir'],
lora_path=model_info['lora_path'],
prompt_template=model_info['prompt_template'],
negative_prompt=model_info['negative_prompt']
)
output_dir = os.path.join(sys.argv[1], model)
os.makedirs(output_dir, exist_ok=True)
# concept test, with control and inversion
input_dir = 'resources/prompts'
for dir, folders, files in os.walk(input_dir):
for file in files:
input_file = os.path.join(dir, file)
file_output_dir = os.path.join(output_dir, file)
print(f'input_file: {input_file}')
print(f'file_output_dir: {file_output_dir}')
face_img_path = os.path.join(face_img_dir, random.choice(faces))
experiment(anything_facemaker, input_file, face_img_path, output_dir=file_output_dir,
controlnet_conditioning_scale=controlnet_conditioning_scale,
strength=strength)
# # concept, with control and inversion
# experiment(anything_facemaker, 'resources/concepts.txt', face_img_path, output_dir='results/concepts/control_inversion',
# controlnet_conditioning_scale=controlnet_conditioning_scale,
# strength=strength)
# # concept test, no control no inversion
# experiment(anything_facemaker, 'resources/concepts_test.txt', face_img_path, output_dir='results/concepts_test/generate',
# controlnet_conditioning_scale=None,
# strength=strength)
# # concept, no control no inversion
# experiment(anything_facemaker, 'resources/concepts.txt', face_img_path, output_dir='results/concepts/generate',
# controlnet_conditioning_scale=None,
# strength=strength) |