Spaces:
Sleeping
Sleeping
add new beit model
Browse files
app.py
CHANGED
@@ -120,6 +120,13 @@ def get_model(model_name):
|
|
120 |
embedding_depth = 2,
|
121 |
n_classes = n_classes)
|
122 |
model.load_weights('model_classification/fossil-new.h5')
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
123 |
else:
|
124 |
raise ValueError(f"Model name '{model_name}' is not recognized")
|
125 |
return model,n_classes
|
@@ -151,6 +158,11 @@ def classify_image(input_image, model_name):
|
|
151 |
model,n_classes = get_model(model_name)
|
152 |
result = inference_resnet_finer_beit(input_image,model,size=384,n_classes=n_classes)
|
153 |
return result
|
|
|
|
|
|
|
|
|
|
|
154 |
return None
|
155 |
|
156 |
def get_embeddings(input_image,model_name):
|
@@ -174,6 +186,11 @@ def get_embeddings(input_image,model_name):
|
|
174 |
model,n_classes = get_model(model_name)
|
175 |
result = inference_resnet_embedding_beit(input_image,model,size=384,n_classes=n_classes)
|
176 |
return result
|
|
|
|
|
|
|
|
|
|
|
177 |
return None
|
178 |
|
179 |
|
@@ -301,9 +318,9 @@ with gr.Blocks(theme='sudeepshouche/minimalist') as demo:
|
|
301 |
|
302 |
with gr.Column():
|
303 |
model_name = gr.Dropdown(
|
304 |
-
["Mummified 170", "Rock 170","Fossils 142","Fossils new"],
|
305 |
multiselect=False,
|
306 |
-
value="Fossils
|
307 |
label="Model",
|
308 |
interactive=True,
|
309 |
info="Choose the model you'd like to use"
|
|
|
120 |
embedding_depth = 2,
|
121 |
n_classes = n_classes)
|
122 |
model.load_weights('model_classification/fossil-new.h5')
|
123 |
+
elif model_name == 'Fossils':
|
124 |
+
n_classes = 142
|
125 |
+
model = get_triplet_model_beit(input_shape = (384, 384, 3),
|
126 |
+
embedding_units = 256,
|
127 |
+
embedding_depth = 2,
|
128 |
+
n_classes = n_classes)
|
129 |
+
model.load_weights('model_classification/fossil-model.h5')
|
130 |
else:
|
131 |
raise ValueError(f"Model name '{model_name}' is not recognized")
|
132 |
return model,n_classes
|
|
|
158 |
model,n_classes = get_model(model_name)
|
159 |
result = inference_resnet_finer_beit(input_image,model,size=384,n_classes=n_classes)
|
160 |
return result
|
161 |
+
elif 'Fossils' ==model_name:
|
162 |
+
from inference_beit import inference_resnet_finer_beit
|
163 |
+
model,n_classes = get_model(model_name)
|
164 |
+
result = inference_resnet_finer_beit(input_image,model,size=384,n_classes=n_classes)
|
165 |
+
return result
|
166 |
return None
|
167 |
|
168 |
def get_embeddings(input_image,model_name):
|
|
|
186 |
model,n_classes = get_model(model_name)
|
187 |
result = inference_resnet_embedding_beit(input_image,model,size=384,n_classes=n_classes)
|
188 |
return result
|
189 |
+
elif 'Fossils' ==model_name:
|
190 |
+
from inference_beit import inference_resnet_embedding_beit
|
191 |
+
model,n_classes = get_model(model_name)
|
192 |
+
result = inference_resnet_embedding_beit(input_image,model,size=384,n_classes=n_classes)
|
193 |
+
return result
|
194 |
return None
|
195 |
|
196 |
|
|
|
318 |
|
319 |
with gr.Column():
|
320 |
model_name = gr.Dropdown(
|
321 |
+
["Mummified 170", "Rock 170","Fossils 142","Fossils new","Fossils"],
|
322 |
multiselect=False,
|
323 |
+
value="Fossils", # default option
|
324 |
label="Model",
|
325 |
interactive=True,
|
326 |
info="Choose the model you'd like to use"
|