Spaces:
Running
Running
fix bugs
Browse files- app.py +4 -18
- requirements.txt +3 -1
app.py
CHANGED
|
@@ -5,8 +5,6 @@ import os
|
|
| 5 |
from src.pipeline import ObjectRemovalPipeline
|
| 6 |
from src.utils import visualize_mask
|
| 7 |
|
| 8 |
-
# --- ZeroGPU Compatibility Shim ---
|
| 9 |
-
# Allows code to run on local CPU/GPU without crashing on 'import spaces'
|
| 10 |
try:
|
| 11 |
import spaces
|
| 12 |
except ImportError:
|
|
@@ -17,11 +15,10 @@ except ImportError:
|
|
| 17 |
return func
|
| 18 |
return decorator
|
| 19 |
|
| 20 |
-
# Initialize pipeline
|
| 21 |
pipeline = ObjectRemovalPipeline()
|
| 22 |
|
| 23 |
def ensure_uint8(image):
|
| 24 |
-
"""Normalize image to uint8 (0-255)"""
|
| 25 |
if image is None: return None
|
| 26 |
image = np.array(image)
|
| 27 |
if image.dtype != np.uint8:
|
|
@@ -31,11 +28,9 @@ def ensure_uint8(image):
|
|
| 31 |
|
| 32 |
@spaces.GPU(duration=120)
|
| 33 |
def step1_detect(image, text_query):
|
| 34 |
-
"""Detect objects and return candidates for user selection"""
|
| 35 |
if image is None or not text_query:
|
| 36 |
return [], [], "Please upload image and enter text."
|
| 37 |
|
| 38 |
-
# 1. Detect & Rank candidates via Pipeline
|
| 39 |
candidates, msg = pipeline.get_candidates(image, text_query)
|
| 40 |
|
| 41 |
if not candidates:
|
|
@@ -43,7 +38,6 @@ def step1_detect(image, text_query):
|
|
| 43 |
|
| 44 |
masks = [c['mask'] for c in candidates]
|
| 45 |
|
| 46 |
-
# 2. Visualize masks for Gallery
|
| 47 |
gallery_imgs = []
|
| 48 |
for i, mask in enumerate(masks):
|
| 49 |
viz = visualize_mask(image, mask)
|
|
@@ -54,30 +48,25 @@ def step1_detect(image, text_query):
|
|
| 54 |
return masks, gallery_imgs, "Select the best match below."
|
| 55 |
|
| 56 |
def on_select(evt: gr.SelectData):
|
| 57 |
-
"""Capture user selection from Gallery"""
|
| 58 |
return evt.index
|
| 59 |
|
| 60 |
@spaces.GPU(duration=120)
|
| 61 |
def step2_remove(image, masks, selected_idx, prompt, shadow_exp):
|
| 62 |
-
"""Inpaint the selected mask"""
|
| 63 |
if not masks or selected_idx is None:
|
| 64 |
return None, "Please select an object first."
|
| 65 |
|
| 66 |
target_mask = masks[selected_idx]
|
| 67 |
|
| 68 |
-
# 3. Inpaint with Shadow Fix logic
|
| 69 |
result = pipeline.inpaint_selected(image, target_mask, prompt, shadow_expansion=shadow_exp)
|
| 70 |
|
| 71 |
return ensure_uint8(result), "Success!"
|
| 72 |
|
| 73 |
-
# CSS for better layout and full image visibility in Gallery
|
| 74 |
css = """
|
| 75 |
.gradio-container {min-height: 0px !important}
|
| 76 |
button.gallery-item {object-fit: contain !important}
|
| 77 |
"""
|
| 78 |
|
| 79 |
-
with gr.Blocks(title="TextEraser"
|
| 80 |
-
# State to hold masks between steps
|
| 81 |
mask_state = gr.State([])
|
| 82 |
idx_state = gr.State(0)
|
| 83 |
|
|
@@ -90,7 +79,6 @@ with gr.Blocks(title="TextEraser", css=css, theme=gr.themes.Soft()) as demo:
|
|
| 90 |
btn_detect = gr.Button("1. Detect Objects", variant="primary")
|
| 91 |
|
| 92 |
with gr.Column(scale=1):
|
| 93 |
-
# Interactive Gallery (Adaptable size)
|
| 94 |
gallery = gr.Gallery(
|
| 95 |
label="Candidates (Select One)",
|
| 96 |
columns=2,
|
|
@@ -109,7 +97,6 @@ with gr.Blocks(title="TextEraser", css=css, theme=gr.themes.Soft()) as demo:
|
|
| 109 |
with gr.Column(scale=1):
|
| 110 |
output_image = gr.Image(label="Final Result", height=400)
|
| 111 |
|
| 112 |
-
# Event Wiring
|
| 113 |
btn_detect.click(
|
| 114 |
fn=step1_detect,
|
| 115 |
inputs=[input_image, text_query],
|
|
@@ -126,8 +113,7 @@ with gr.Blocks(title="TextEraser", css=css, theme=gr.themes.Soft()) as demo:
|
|
| 126 |
|
| 127 |
if __name__ == "__main__":
|
| 128 |
parser = argparse.ArgumentParser()
|
| 129 |
-
parser.add_argument("--share", action="store_true"
|
| 130 |
args = parser.parse_args()
|
| 131 |
|
| 132 |
-
|
| 133 |
-
demo.queue().launch(share=args.share)
|
|
|
|
| 5 |
from src.pipeline import ObjectRemovalPipeline
|
| 6 |
from src.utils import visualize_mask
|
| 7 |
|
|
|
|
|
|
|
| 8 |
try:
|
| 9 |
import spaces
|
| 10 |
except ImportError:
|
|
|
|
| 15 |
return func
|
| 16 |
return decorator
|
| 17 |
|
| 18 |
+
# Initialize pipeline
|
| 19 |
pipeline = ObjectRemovalPipeline()
|
| 20 |
|
| 21 |
def ensure_uint8(image):
|
|
|
|
| 22 |
if image is None: return None
|
| 23 |
image = np.array(image)
|
| 24 |
if image.dtype != np.uint8:
|
|
|
|
| 28 |
|
| 29 |
@spaces.GPU(duration=120)
|
| 30 |
def step1_detect(image, text_query):
|
|
|
|
| 31 |
if image is None or not text_query:
|
| 32 |
return [], [], "Please upload image and enter text."
|
| 33 |
|
|
|
|
| 34 |
candidates, msg = pipeline.get_candidates(image, text_query)
|
| 35 |
|
| 36 |
if not candidates:
|
|
|
|
| 38 |
|
| 39 |
masks = [c['mask'] for c in candidates]
|
| 40 |
|
|
|
|
| 41 |
gallery_imgs = []
|
| 42 |
for i, mask in enumerate(masks):
|
| 43 |
viz = visualize_mask(image, mask)
|
|
|
|
| 48 |
return masks, gallery_imgs, "Select the best match below."
|
| 49 |
|
| 50 |
def on_select(evt: gr.SelectData):
|
|
|
|
| 51 |
return evt.index
|
| 52 |
|
| 53 |
@spaces.GPU(duration=120)
|
| 54 |
def step2_remove(image, masks, selected_idx, prompt, shadow_exp):
|
|
|
|
| 55 |
if not masks or selected_idx is None:
|
| 56 |
return None, "Please select an object first."
|
| 57 |
|
| 58 |
target_mask = masks[selected_idx]
|
| 59 |
|
|
|
|
| 60 |
result = pipeline.inpaint_selected(image, target_mask, prompt, shadow_expansion=shadow_exp)
|
| 61 |
|
| 62 |
return ensure_uint8(result), "Success!"
|
| 63 |
|
|
|
|
| 64 |
css = """
|
| 65 |
.gradio-container {min-height: 0px !important}
|
| 66 |
button.gallery-item {object-fit: contain !important}
|
| 67 |
"""
|
| 68 |
|
| 69 |
+
with gr.Blocks(title="TextEraser") as demo:
|
|
|
|
| 70 |
mask_state = gr.State([])
|
| 71 |
idx_state = gr.State(0)
|
| 72 |
|
|
|
|
| 79 |
btn_detect = gr.Button("1. Detect Objects", variant="primary")
|
| 80 |
|
| 81 |
with gr.Column(scale=1):
|
|
|
|
| 82 |
gallery = gr.Gallery(
|
| 83 |
label="Candidates (Select One)",
|
| 84 |
columns=2,
|
|
|
|
| 97 |
with gr.Column(scale=1):
|
| 98 |
output_image = gr.Image(label="Final Result", height=400)
|
| 99 |
|
|
|
|
| 100 |
btn_detect.click(
|
| 101 |
fn=step1_detect,
|
| 102 |
inputs=[input_image, text_query],
|
|
|
|
| 113 |
|
| 114 |
if __name__ == "__main__":
|
| 115 |
parser = argparse.ArgumentParser()
|
| 116 |
+
parser.add_argument("--share", action="store_true")
|
| 117 |
args = parser.parse_args()
|
| 118 |
|
| 119 |
+
demo.queue().launch(share=args.share, css=css)
|
|
|
requirements.txt
CHANGED
|
@@ -19,4 +19,6 @@ Pillow
|
|
| 19 |
sniffio
|
| 20 |
spaces
|
| 21 |
clip
|
| 22 |
-
git+https://github.com/facebookresearch/sam2.git
|
|
|
|
|
|
|
|
|
| 19 |
sniffio
|
| 20 |
spaces
|
| 21 |
clip
|
| 22 |
+
git+https://github.com/facebookresearch/sam2.git
|
| 23 |
+
accelerate
|
| 24 |
+
ftfy
|