Spaces:
Running
Running
Commit
·
dd559e1
1
Parent(s):
46af7ed
WIP
Browse files
app.py
CHANGED
@@ -4,6 +4,7 @@ import spaces
|
|
4 |
import os
|
5 |
import random
|
6 |
|
|
|
7 |
import torch
|
8 |
from PIL import Image
|
9 |
import cv2
|
@@ -32,6 +33,8 @@ Before running, set the `HUGGINGFACE_TOKEN` environment variable **or** call
|
|
32 |
`login("<YOUR_HF_TOKEN>")` explicitly.
|
33 |
"""
|
34 |
|
|
|
|
|
35 |
# --------------------------------------------------
|
36 |
# Model & pipeline setup
|
37 |
# --------------------------------------------------
|
@@ -76,16 +79,23 @@ MAX_SEED = 100
|
|
76 |
# --------------------------------------------------
|
77 |
|
78 |
|
79 |
-
def _preview_canny(
|
|
|
|
|
80 |
arr = np.array(pil_img.convert("RGB"))
|
81 |
-
edges = cv2.Canny(arr,
|
82 |
edges_rgb = cv2.cvtColor(edges, cv2.COLOR_GRAY2RGB)
|
83 |
return Image.fromarray(edges_rgb)
|
84 |
|
85 |
|
86 |
-
def _make_preview(
|
|
|
|
|
|
|
|
|
|
|
87 |
if mode == "canny":
|
88 |
-
return _preview_canny(control_image)
|
89 |
# For other modes you can plug in your own visualiser later
|
90 |
return control_image
|
91 |
|
@@ -105,6 +115,8 @@ def infer(
|
|
105 |
randomize_seed: bool,
|
106 |
guidance_scale: float,
|
107 |
num_inference_steps: int,
|
|
|
|
|
108 |
):
|
109 |
if control_image is None:
|
110 |
raise gr.Error("Please upload a control image first.")
|
@@ -115,8 +127,12 @@ def infer(
|
|
115 |
gen = torch.Generator(device).manual_seed(seed)
|
116 |
w, h = control_image.size
|
117 |
|
|
|
|
|
|
|
|
|
118 |
result = pipe(
|
119 |
-
prompt=
|
120 |
control_image=[control_image],
|
121 |
control_mode=[MODE_MAPPING[mode]],
|
122 |
width=w,
|
@@ -127,8 +143,7 @@ def infer(
|
|
127 |
generator=gen,
|
128 |
).images[0]
|
129 |
|
130 |
-
|
131 |
-
return result, seed, preview
|
132 |
|
133 |
|
134 |
# --------------------------------------------------
|
@@ -148,23 +163,23 @@ with gr.Blocks(css=css, elem_id="wrapper") as demo:
|
|
148 |
control_image = gr.Image(
|
149 |
label="Upload a processed control image",
|
150 |
type="pil",
|
151 |
-
height=512,
|
152 |
)
|
153 |
-
result_image = gr.Image(label="Result", height=512)
|
154 |
-
preview_image = gr.Image(label="Pre‑processed Cond", height=512)
|
155 |
|
156 |
# ------------ Prompt ------------
|
157 |
-
prompt_txt = gr.Textbox(label="Prompt", value="
|
158 |
|
159 |
# ------------ ControlNet settings ------------
|
160 |
with gr.Row():
|
161 |
with gr.Column():
|
162 |
gr.Markdown("### ControlNet")
|
163 |
mode_radio = gr.Radio(
|
164 |
-
choices=list(MODE_MAPPING.keys()), value="
|
165 |
)
|
166 |
strength_slider = gr.Slider(
|
167 |
-
0.0, 1.0, value=0.
|
168 |
)
|
169 |
with gr.Column():
|
170 |
seed_slider = gr.Slider(0, MAX_SEED, step=1, value=42, label="Seed")
|
@@ -174,6 +189,15 @@ with gr.Blocks(css=css, elem_id="wrapper") as demo:
|
|
174 |
)
|
175 |
steps_slider = gr.Slider(1, 50, step=1, value=24, label="Inference steps")
|
176 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
177 |
submit_btn = gr.Button("Submit")
|
178 |
|
179 |
submit_btn.click(
|
@@ -187,6 +211,8 @@ with gr.Blocks(css=css, elem_id="wrapper") as demo:
|
|
187 |
randomize_chk,
|
188 |
guidance_slider,
|
189 |
steps_slider,
|
|
|
|
|
190 |
],
|
191 |
outputs=[result_image, seed_slider, preview_image],
|
192 |
)
|
|
|
4 |
import os
|
5 |
import random
|
6 |
|
7 |
+
import subprocess
|
8 |
import torch
|
9 |
from PIL import Image
|
10 |
import cv2
|
|
|
33 |
`login("<YOUR_HF_TOKEN>")` explicitly.
|
34 |
"""
|
35 |
|
36 |
+
subprocess.run("rm -rf /data-nvme/zerogpu-offload/*", env={}, shell=True)
|
37 |
+
|
38 |
# --------------------------------------------------
|
39 |
# Model & pipeline setup
|
40 |
# --------------------------------------------------
|
|
|
79 |
# --------------------------------------------------
|
80 |
|
81 |
|
82 |
+
def _preview_canny(
|
83 |
+
pil_img: Image.Image, canny_threshold_1: int, canny_threshold_2: int
|
84 |
+
) -> Image.Image:
|
85 |
arr = np.array(pil_img.convert("RGB"))
|
86 |
+
edges = cv2.Canny(arr, threshold1=canny_threshold_1, threshold2=canny_threshold_2)
|
87 |
edges_rgb = cv2.cvtColor(edges, cv2.COLOR_GRAY2RGB)
|
88 |
return Image.fromarray(edges_rgb)
|
89 |
|
90 |
|
91 |
+
def _make_preview(
|
92 |
+
control_image: Image.Image,
|
93 |
+
mode: str,
|
94 |
+
canny_threshold_1: int,
|
95 |
+
canny_threshold_2: int,
|
96 |
+
) -> Image.Image:
|
97 |
if mode == "canny":
|
98 |
+
return _preview_canny(control_image, canny_threshold_1, canny_threshold_2)
|
99 |
# For other modes you can plug in your own visualiser later
|
100 |
return control_image
|
101 |
|
|
|
115 |
randomize_seed: bool,
|
116 |
guidance_scale: float,
|
117 |
num_inference_steps: int,
|
118 |
+
canny_threshold_1: int,
|
119 |
+
canny_threshold_2: int,
|
120 |
):
|
121 |
if control_image is None:
|
122 |
raise gr.Error("Please upload a control image first.")
|
|
|
127 |
gen = torch.Generator(device).manual_seed(seed)
|
128 |
w, h = control_image.size
|
129 |
|
130 |
+
preprocessed = _make_preview(
|
131 |
+
control_image, mode, canny_threshold_1, canny_threshold_2
|
132 |
+
)
|
133 |
+
|
134 |
result = pipe(
|
135 |
+
prompt=preprocessed,
|
136 |
control_image=[control_image],
|
137 |
control_mode=[MODE_MAPPING[mode]],
|
138 |
width=w,
|
|
|
143 |
generator=gen,
|
144 |
).images[0]
|
145 |
|
146 |
+
return result, seed, preprocessed
|
|
|
147 |
|
148 |
|
149 |
# --------------------------------------------------
|
|
|
163 |
control_image = gr.Image(
|
164 |
label="Upload a processed control image",
|
165 |
type="pil",
|
166 |
+
height=512 + 256,
|
167 |
)
|
168 |
+
result_image = gr.Image(label="Result", height=512 + 256)
|
169 |
+
preview_image = gr.Image(label="Pre‑processed Cond", height=512 + 256)
|
170 |
|
171 |
# ------------ Prompt ------------
|
172 |
+
prompt_txt = gr.Textbox(label="Prompt", value="A beautiful image", lines=1)
|
173 |
|
174 |
# ------------ ControlNet settings ------------
|
175 |
with gr.Row():
|
176 |
with gr.Column():
|
177 |
gr.Markdown("### ControlNet")
|
178 |
mode_radio = gr.Radio(
|
179 |
+
choices=list(MODE_MAPPING.keys()), value="canny", label="Mode"
|
180 |
)
|
181 |
strength_slider = gr.Slider(
|
182 |
+
0.0, 1.0, value=0.8, step=0.01, label="control strength"
|
183 |
)
|
184 |
with gr.Column():
|
185 |
seed_slider = gr.Slider(0, MAX_SEED, step=1, value=42, label="Seed")
|
|
|
189 |
)
|
190 |
steps_slider = gr.Slider(1, 50, step=1, value=24, label="Inference steps")
|
191 |
|
192 |
+
with gr.Row():
|
193 |
+
with gr.Column():
|
194 |
+
gr.Markdown("### Preprocess")
|
195 |
+
canny_threshold_1 = gr.Slider(
|
196 |
+
0, 500, step=1, value=100, label="Canny threshold 1"
|
197 |
+
)
|
198 |
+
canny_threshold_2 = gr.Slider(
|
199 |
+
0, 500, step=1, value=200, label="Canny threshold 2"
|
200 |
+
)
|
201 |
submit_btn = gr.Button("Submit")
|
202 |
|
203 |
submit_btn.click(
|
|
|
211 |
randomize_chk,
|
212 |
guidance_slider,
|
213 |
steps_slider,
|
214 |
+
canny_threshold_1,
|
215 |
+
canny_threshold_2,
|
216 |
],
|
217 |
outputs=[result_image, seed_slider, preview_image],
|
218 |
)
|