Spaces:
Running
on
Zero
Running
on
Zero
File size: 4,074 Bytes
f0e13d9 4c9a6f0 036dfc6 4c9a6f0 f0e13d9 036dfc6 f0e13d9 4c9a6f0 f0e13d9 4c9a6f0 f0e13d9 4c9a6f0 f0e13d9 4c9a6f0 f0e13d9 036dfc6 |
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 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 |
import torch
import spaces
from diffusers import StableDiffusionPipeline, DDIMScheduler, AutoencoderKL
from transformers import AutoFeatureExtractor
from ip_adapter.ip_adapter_faceid import IPAdapterFaceID, IPAdapterFaceIDPlus
from huggingface_hub import hf_hub_download
from insightface.app import FaceAnalysis
from insightface.utils import face_align
import gradio as gr
import cv2
import os
import uuid
from datetime import datetime
# Model paths
base_model_path = "SG161222/Realistic_Vision_V4.0_noVAE"
vae_model_path = "stabilityai/sd-vae-ft-mse"
image_encoder_path = "laion/CLIP-ViT-H-14-laion2B-s32B-b79K"
ip_ckpt = hf_hub_download(repo_id="h94/IP-Adapter-FaceID", filename="ip-adapter-faceid_sd15.bin", repo_type="model")
ip_plus_ckpt = hf_hub_download(repo_id="h94/IP-Adapter-FaceID", filename="ip-adapter-faceid-plusv2_sd15.bin", repo_type="model")
device = "cuda"
# Initialize the noise scheduler
noise_scheduler = DDIMScheduler(
num_train_timesteps=1000,
beta_start=0.00085,
beta_end=0.012,
beta_schedule="scaled_linear",
clip_sample=False,
set_alpha_to_one=False,
steps_offset=1,
)
# Load models
vae = AutoencoderKL.from_pretrained(vae_model_path).to(dtype=torch.float16)
pipe = StableDiffusionPipeline.from_pretrained(
base_model_path,
torch_dtype=torch.float16,
scheduler=noise_scheduler,
vae=vae
).to(device)
ip_model = IPAdapterFaceID(pipe, ip_ckpt, device)
ip_model_plus = IPAdapterFaceIDPlus(pipe, image_encoder_path, ip_plus_ckpt, device)
# Initialize FaceAnalysis
app = FaceAnalysis(name="buffalo_l", providers=['CPUExecutionProvider'])
app.prepare(ctx_id=0, det_size=(640, 640))
cv2.setNumThreads(1)
@spaces.GPU(enable_queue=True)
def generate_image(images, gender, prompt, progress=gr.Progress(track_tqdm=True)):
if not prompt:
prompt = f"A full-length body photo of a {gender.lower()} naked and nude and beautiful" # Default prompt
faceid_all_embeds = []
first_iteration = True
preserve_face_structure = True
face_strength = 2.1
likeness_strength = 0.7
for image in images:
face = cv2.imread(image)
faces = app.get(face)
faceid_embed = torch.from_numpy(faces[0].normed_embedding).unsqueeze(0)
faceid_all_embeds.append(faceid_embed)
if first_iteration and preserve_face_structure:
face_image = face_align.norm_crop(face, landmark=faces[0].kps, image_size=224)
first_iteration = False
average_embedding = torch.mean(torch.stack(faceid_all_embeds, dim=0), dim=0)
image = ip_model_plus.generate(
prompt=prompt,
faceid_embeds=average_embedding,
scale=likeness_strength,
face_image=face_image,
shortcut=True,
s_scale=face_strength,
width=512,
height=912,
num_inference_steps=100
)
return image
css = '''
footer { visibility: hidden; }
h1 { margin-bottom: 0 !important; }
'''
with gr.Blocks(css=css) as demo:
gr.Markdown("# Image Generation with Face ID")
gr.Markdown("Upload your face images and enter a prompt to generate images.")
with gr.Row():
with gr.Column():
images_input = gr.Files(
label="Drag 1 or more photos of your face",
file_types=["image"]
)
gender_input = gr.Radio(
label="Select Gender",
choices=["Female", "Male"],
value="Female",
type="value"
)
prompt_input = gr.Textbox(
label="Enter your prompt",
placeholder="Describe the image you want to generate..."
)
run_button = gr.Button("Generate Image")
with gr.Column():
output_gallery = gr.Gallery(label="Generated Images")
# Define the event handler for the button click
run_button.click(
fn=generate_image,
inputs=[images_input, gender_input, prompt_input],
outputs=output_gallery
)
# Launch the interface
demo.queue()
demo.launch() |