Spaces:
Running
Running
Update app.py
#1
by
Jhony23German
- opened
app.py
CHANGED
|
@@ -1,151 +1,170 @@
|
|
| 1 |
-
import hashlib
|
| 2 |
-
import os
|
| 3 |
-
from io import BytesIO
|
| 4 |
-
|
| 5 |
-
import gradio as gr
|
| 6 |
-
import grpc
|
| 7 |
-
from PIL import Image
|
| 8 |
-
from cachetools import LRUCache
|
| 9 |
-
|
| 10 |
-
from inference_pb2 import HairSwapRequest, HairSwapResponse
|
| 11 |
-
from inference_pb2_grpc import HairSwapServiceStub
|
| 12 |
-
from utils.shape_predictor import align_face
|
| 13 |
-
|
| 14 |
-
|
| 15 |
-
def get_bytes(img):
|
| 16 |
-
if img is None:
|
| 17 |
-
return img
|
| 18 |
-
|
| 19 |
-
buffered = BytesIO()
|
| 20 |
-
img.save(buffered, format="JPEG")
|
| 21 |
-
return buffered.getvalue()
|
| 22 |
-
|
| 23 |
-
|
| 24 |
-
def bytes_to_image(image: bytes) -> Image.Image:
|
| 25 |
-
image = Image.open(BytesIO(image))
|
| 26 |
-
return image
|
| 27 |
-
|
| 28 |
-
|
| 29 |
-
def center_crop(img):
|
| 30 |
-
width, height = img.size
|
| 31 |
-
side = min(width, height)
|
| 32 |
-
|
| 33 |
-
left = (width - side) / 2
|
| 34 |
-
top = (height - side) / 2
|
| 35 |
-
right = (width + side) / 2
|
| 36 |
-
bottom = (height + side) / 2
|
| 37 |
-
|
| 38 |
-
img = img.crop((left, top, right, bottom))
|
| 39 |
-
return img
|
| 40 |
-
|
| 41 |
-
|
| 42 |
-
def resize(name):
|
| 43 |
-
def resize_inner(img, align):
|
| 44 |
-
global align_cache
|
| 45 |
-
|
| 46 |
-
if name in align:
|
| 47 |
-
img_hash = hashlib.md5(get_bytes(img)).hexdigest()
|
| 48 |
-
|
| 49 |
-
if img_hash not in align_cache:
|
| 50 |
-
img = align_face(img, return_tensors=False)[0]
|
| 51 |
-
align_cache[img_hash] = img
|
| 52 |
-
else:
|
| 53 |
-
img = align_cache[img_hash]
|
| 54 |
-
|
| 55 |
-
elif img.size != (1024, 1024):
|
| 56 |
-
img = center_crop(img)
|
| 57 |
-
img = img.resize((1024, 1024), Image.Resampling.LANCZOS)
|
| 58 |
-
|
| 59 |
-
return img
|
| 60 |
-
|
| 61 |
-
return resize_inner
|
| 62 |
-
|
| 63 |
-
|
| 64 |
-
def swap_hair(face, shape, color, blending, poisson_iters, poisson_erosion):
|
| 65 |
-
|
| 66 |
-
|
| 67 |
-
|
| 68 |
-
|
| 69 |
-
elif not
|
| 70 |
-
return gr.update(visible=False),
|
| 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 |
-
|
| 129 |
-
|
| 130 |
-
|
| 131 |
-
|
| 132 |
-
|
| 133 |
-
|
| 134 |
-
|
| 135 |
-
|
| 136 |
-
|
| 137 |
-
|
| 138 |
-
|
| 139 |
-
|
| 140 |
-
|
| 141 |
-
|
| 142 |
-
|
| 143 |
-
|
| 144 |
-
|
| 145 |
-
|
| 146 |
-
|
| 147 |
-
|
| 148 |
-
|
| 149 |
-
|
| 150 |
-
|
| 151 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
import hashlib
|
| 2 |
+
import os
|
| 3 |
+
from io import BytesIO
|
| 4 |
+
|
| 5 |
+
import gradio as gr
|
| 6 |
+
import grpc
|
| 7 |
+
from PIL import Image
|
| 8 |
+
from cachetools import LRUCache
|
| 9 |
+
|
| 10 |
+
from inference_pb2 import HairSwapRequest, HairSwapResponse
|
| 11 |
+
from inference_pb2_grpc import HairSwapServiceStub
|
| 12 |
+
from utils.shape_predictor import align_face
|
| 13 |
+
|
| 14 |
+
|
| 15 |
+
def get_bytes(img):
|
| 16 |
+
if img is None:
|
| 17 |
+
return img
|
| 18 |
+
|
| 19 |
+
buffered = BytesIO()
|
| 20 |
+
img.save(buffered, format="JPEG")
|
| 21 |
+
return buffered.getvalue()
|
| 22 |
+
|
| 23 |
+
|
| 24 |
+
def bytes_to_image(image: bytes) -> Image.Image:
|
| 25 |
+
image = Image.open(BytesIO(image))
|
| 26 |
+
return image
|
| 27 |
+
|
| 28 |
+
|
| 29 |
+
def center_crop(img):
|
| 30 |
+
width, height = img.size
|
| 31 |
+
side = min(width, height)
|
| 32 |
+
|
| 33 |
+
left = (width - side) / 2
|
| 34 |
+
top = (height - side) / 2
|
| 35 |
+
right = (width + side) / 2
|
| 36 |
+
bottom = (height + side) / 2
|
| 37 |
+
|
| 38 |
+
img = img.crop((left, top, right, bottom))
|
| 39 |
+
return img
|
| 40 |
+
|
| 41 |
+
|
| 42 |
+
def resize(name):
|
| 43 |
+
def resize_inner(img, align):
|
| 44 |
+
global align_cache
|
| 45 |
+
|
| 46 |
+
if name in align:
|
| 47 |
+
img_hash = hashlib.md5(get_bytes(img)).hexdigest()
|
| 48 |
+
|
| 49 |
+
if img_hash not in align_cache:
|
| 50 |
+
img = align_face(img, return_tensors=False)[0]
|
| 51 |
+
align_cache[img_hash] = img
|
| 52 |
+
else:
|
| 53 |
+
img = align_cache[img_hash]
|
| 54 |
+
|
| 55 |
+
elif img.size != (1024, 1024):
|
| 56 |
+
img = center_crop(img)
|
| 57 |
+
img = img.resize((1024, 1024), Image.Resampling.LANCZOS)
|
| 58 |
+
|
| 59 |
+
return img
|
| 60 |
+
|
| 61 |
+
return resize_inner
|
| 62 |
+
|
| 63 |
+
|
| 64 |
+
def swap_hair(face, shape, color, blending, poisson_iters, poisson_erosion):
|
| 65 |
+
# Verificar que al menos uno de shape o color esté presente junto con face
|
| 66 |
+
if not face and not (shape or color):
|
| 67 |
+
return (gr.update(visible=False),
|
| 68 |
+
gr.update(value="Need to upload a face and at least a shape or color ❗", visible=True))
|
| 69 |
+
elif not face:
|
| 70 |
+
return (gr.update(visible=False),
|
| 71 |
+
gr.update(value="Need to upload a face ❗", visible=True))
|
| 72 |
+
elif not (shape or color):
|
| 73 |
+
return (gr.update(visible=False),
|
| 74 |
+
gr.update(value="Need to upload at least a shape or color ❗", visible=True))
|
| 75 |
+
|
| 76 |
+
# Obtener los bytes de los blobs
|
| 77 |
+
face_bytes, shape_bytes, color_bytes = map(lambda item: get_bytes(item) if item else None, (face, shape, color))
|
| 78 |
+
|
| 79 |
+
# Asignar valores por defecto si no están presentes
|
| 80 |
+
if shape_bytes is None:
|
| 81 |
+
shape_bytes = b'face'
|
| 82 |
+
if color_bytes is None:
|
| 83 |
+
color_bytes = b'shape'
|
| 84 |
+
|
| 85 |
+
try:
|
| 86 |
+
with grpc.insecure_channel(os.environ['SERVER']) as channel:
|
| 87 |
+
stub = HairSwapServiceStub(channel)
|
| 88 |
+
|
| 89 |
+
output: HairSwapResponse = stub.swap(
|
| 90 |
+
HairSwapRequest(face=face_bytes, shape=shape_bytes, color=color_bytes, blending=blending,
|
| 91 |
+
poisson_iters=poisson_iters, poisson_erosion=poisson_erosion, use_cache=True)
|
| 92 |
+
)
|
| 93 |
+
|
| 94 |
+
output_image = bytes_to_image(output.image)
|
| 95 |
+
return (gr.update(value=output_image, visible=True),
|
| 96 |
+
gr.update(visible=False))
|
| 97 |
+
except grpc.RpcError as e:
|
| 98 |
+
# Manejo de errores de gRPC
|
| 99 |
+
error_message = f"gRPC error: {e.code()}: {e.details()}"
|
| 100 |
+
return (gr.update(visible=False),
|
| 101 |
+
gr.update(value=error_message, visible=True))
|
| 102 |
+
except Exception as e:
|
| 103 |
+
# Manejo de cualquier otro error
|
| 104 |
+
error_message = f"Unexpected error: {str(e)}"
|
| 105 |
+
return (gr.update(visible=False),
|
| 106 |
+
gr.update(value=error_message, visible=True))
|
| 107 |
+
|
| 108 |
+
|
| 109 |
+
|
| 110 |
+
def get_demo():
|
| 111 |
+
with gr.Blocks() as demo:
|
| 112 |
+
gr.Markdown("## HairFastGan")
|
| 113 |
+
gr.Markdown(
|
| 114 |
+
'<div style="display: flex; align-items: center; gap: 10px;">'
|
| 115 |
+
'<span>Official HairFastGAN Gradio demo:</span>'
|
| 116 |
+
'<a href="https://arxiv.org/abs/2404.01094"><img src="https://img.shields.io/badge/arXiv-2404.01094-b31b1b.svg" height=22.5></a>'
|
| 117 |
+
'<a href="https://github.com/AIRI-Institute/HairFastGAN"><img src="https://img.shields.io/badge/github-%23121011.svg?style=for-the-badge&logo=github&logoColor=white" height=22.5></a>'
|
| 118 |
+
'<a href="https://huggingface.co/AIRI-Institute/HairFastGAN"><img src="https://huggingface.co/datasets/huggingface/badges/resolve/main/model-on-hf-md.svg" height=22.5></a>'
|
| 119 |
+
'<a href="https://colab.research.google.com/#fileId=https://huggingface.co/AIRI-Institute/HairFastGAN/blob/main/notebooks/HairFast_inference.ipynb"><img src="https://colab.research.google.com/assets/colab-badge.svg" height=22.5></a>'
|
| 120 |
+
'</div>'
|
| 121 |
+
)
|
| 122 |
+
with gr.Row():
|
| 123 |
+
with gr.Column():
|
| 124 |
+
source = gr.Image(label="Source photo to try on the hairstyle", type="pil")
|
| 125 |
+
with gr.Row():
|
| 126 |
+
shape = gr.Image(label="Shape photo with desired hairstyle (optional)", type="pil")
|
| 127 |
+
color = gr.Image(label="Color photo with desired hair color (optional)", type="pil")
|
| 128 |
+
with gr.Accordion("Advanced Options", open=False):
|
| 129 |
+
blending = gr.Radio(["Article", "Alternative_v1", "Alternative_v2"], value='Article',
|
| 130 |
+
label="Color Encoder version", info="Selects a model for hair color transfer.")
|
| 131 |
+
poisson_iters = gr.Slider(0, 2500, value=0, step=1, label="Poisson iters",
|
| 132 |
+
info="The power of blending with the original image, helps to recover more details. Not included in the article, disabled by default.")
|
| 133 |
+
poisson_erosion = gr.Slider(1, 100, value=15, step=1, label="Poisson erosion",
|
| 134 |
+
info="Smooths out the blending area.")
|
| 135 |
+
align = gr.CheckboxGroup(["Face", "Shape", "Color"], value=["Face", "Shape", "Color"],
|
| 136 |
+
label="Image cropping [recommended]",
|
| 137 |
+
info="Selects which images to crop by face")
|
| 138 |
+
btn = gr.Button("Get the haircut")
|
| 139 |
+
with gr.Column():
|
| 140 |
+
output = gr.Image(label="Your result")
|
| 141 |
+
error_message = gr.Textbox(label="⚠️ Error ⚠️", visible=False, elem_classes="error-message")
|
| 142 |
+
|
| 143 |
+
gr.Examples(examples=[["input/0.png", "input/1.png", "input/2.png"], ["input/6.png", "input/7.png", None],
|
| 144 |
+
["input/10.jpg", None, "input/11.jpg"]],
|
| 145 |
+
inputs=[source, shape, color], outputs=output)
|
| 146 |
+
|
| 147 |
+
source.upload(fn=resize('Face'), inputs=[source, align], outputs=source)
|
| 148 |
+
shape.upload(fn=resize('Shape'), inputs=[shape, align], outputs=shape)
|
| 149 |
+
color.upload(fn=resize('Color'), inputs=[color, align], outputs=color)
|
| 150 |
+
|
| 151 |
+
btn.click(fn=swap_hair, inputs=[source, shape, color, blending, poisson_iters, poisson_erosion],
|
| 152 |
+
outputs=[output, error_message])
|
| 153 |
+
|
| 154 |
+
gr.Markdown('''To cite the paper by the authors
|
| 155 |
+
```
|
| 156 |
+
@article{nikolaev2024hairfastgan,
|
| 157 |
+
title={HairFastGAN: Realistic and Robust Hair Transfer with a Fast Encoder-Based Approach},
|
| 158 |
+
author={Nikolaev, Maxim and Kuznetsov, Mikhail and Vetrov, Dmitry and Alanov, Aibek},
|
| 159 |
+
journal={arXiv preprint arXiv:2404.01094},
|
| 160 |
+
year={2024}
|
| 161 |
+
}
|
| 162 |
+
```
|
| 163 |
+
''')
|
| 164 |
+
return demo
|
| 165 |
+
|
| 166 |
+
|
| 167 |
+
if __name__ == '__main__':
|
| 168 |
+
align_cache = LRUCache(maxsize=10)
|
| 169 |
+
demo = get_demo()
|
| 170 |
+
demo.launch(server_name="0.0.0.0", server_port=7860)
|