major UI overhaul
Browse files
app.py
CHANGED
@@ -128,23 +128,33 @@ allowed_tags = list(tags.keys())
|
|
128 |
for idx, tag in enumerate(allowed_tags):
|
129 |
allowed_tags[idx] = tag.replace("_", " ")
|
130 |
|
|
|
|
|
131 |
@spaces.GPU(duration=5)
|
132 |
-
def
|
|
|
133 |
img = image.convert('RGB')
|
134 |
tensor = transform(img).unsqueeze(0)
|
135 |
|
136 |
with torch.no_grad():
|
137 |
logits = model(tensor)
|
138 |
probabilities = torch.nn.functional.sigmoid(logits[0])
|
139 |
-
indices = torch.where(probabilities
|
140 |
values = probabilities[indices]
|
141 |
|
142 |
tag_score = dict()
|
143 |
for i in range(indices.size(0)):
|
144 |
tag_score[allowed_tags[indices[i]]] = values[i].item()
|
145 |
sorted_tag_score = dict(sorted(tag_score.items(), key=lambda item: item[1], reverse=True))
|
146 |
-
|
147 |
-
return
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
148 |
|
149 |
with gr.Blocks(css=".output-class { display: none; }") as demo:
|
150 |
gr.Markdown("""
|
@@ -153,14 +163,24 @@ with gr.Blocks(css=".output-class { display: none; }") as demo:
|
|
153 |
|
154 |
This tagger is the result of joint efforts between members of the RedRocket team. Special thanks to Minotoro at frosting.ai for providing the compute power for this project.
|
155 |
""")
|
156 |
-
gr.
|
157 |
-
|
158 |
-
|
159 |
-
|
160 |
-
|
161 |
-
gr.
|
162 |
-
|
163 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
164 |
)
|
165 |
|
166 |
if __name__ == "__main__":
|
|
|
128 |
for idx, tag in enumerate(allowed_tags):
|
129 |
allowed_tags[idx] = tag.replace("_", " ")
|
130 |
|
131 |
+
sorted_tag_score = {}
|
132 |
+
|
133 |
@spaces.GPU(duration=5)
|
134 |
+
def run_classifier(image, threshold):
|
135 |
+
global sorted_tag_score
|
136 |
img = image.convert('RGB')
|
137 |
tensor = transform(img).unsqueeze(0)
|
138 |
|
139 |
with torch.no_grad():
|
140 |
logits = model(tensor)
|
141 |
probabilities = torch.nn.functional.sigmoid(logits[0])
|
142 |
+
indices = torch.where(probabilities, 250).indices
|
143 |
values = probabilities[indices]
|
144 |
|
145 |
tag_score = dict()
|
146 |
for i in range(indices.size(0)):
|
147 |
tag_score[allowed_tags[indices[i]]] = values[i].item()
|
148 |
sorted_tag_score = dict(sorted(tag_score.items(), key=lambda item: item[1], reverse=True))
|
149 |
+
|
150 |
+
return create_tags(threshold)
|
151 |
+
|
152 |
+
def create_tags(threshold):
|
153 |
+
global sorted_tag_score
|
154 |
+
filtered_tag_score = {key: value for key, value in sorted_tag_score.items() if value > threshold}
|
155 |
+
text_no_impl = ", ".join(filtered_tag_score.keys())
|
156 |
+
return text_no_impl, filtered_tag_score
|
157 |
+
|
158 |
|
159 |
with gr.Blocks(css=".output-class { display: none; }") as demo:
|
160 |
gr.Markdown("""
|
|
|
163 |
|
164 |
This tagger is the result of joint efforts between members of the RedRocket team. Special thanks to Minotoro at frosting.ai for providing the compute power for this project.
|
165 |
""")
|
166 |
+
with gr.Row():
|
167 |
+
with gr.Column():
|
168 |
+
image_input = gr.Image(label="Source", sources=['upload'], type='pil', height="60vh", show_label=False)
|
169 |
+
threshold_slider = gr.Slider(minimum=0.00, maximum=1.00, step=0.01, value=0.20, label="Threshold")
|
170 |
+
with gr.Column():
|
171 |
+
tag_string = gr.Textbox(label="Tag String")
|
172 |
+
label_box = gr.Label(label="Tag Predictions", num_top_classes=250, show_label=False)
|
173 |
+
|
174 |
+
image_input.upload(
|
175 |
+
fn=run_classifier,
|
176 |
+
inputs=[image_input, threshold_slider],
|
177 |
+
outputs=[tag_string, label_box]
|
178 |
+
)
|
179 |
+
|
180 |
+
threshold_slider.input(
|
181 |
+
fn=create_tags,
|
182 |
+
inputs=[threshold_slider],
|
183 |
+
outputs=[tag_string, label_box]
|
184 |
)
|
185 |
|
186 |
if __name__ == "__main__":
|