Spaces:
Sleeping
Sleeping
adding closest samples
Browse files- app.py +81 -23
- inference_resnet.py +15 -14
- inference_sam.py +5 -4
- pre-requirements.txt +2 -1
app.py
CHANGED
@@ -9,6 +9,7 @@ if os.getenv('SYSTEM') == 'spaces':
|
|
9 |
subprocess.call('pip install git+https://github.com/cocodataset/panopticapi.git'.split())
|
10 |
subprocess.call('pip install python-dotenv'.split())
|
11 |
subprocess.call('pip install torch torchvision '.split())
|
|
|
12 |
|
13 |
import gradio as gr
|
14 |
from huggingface_hub import snapshot_download
|
@@ -19,13 +20,38 @@ import numpy as np
|
|
19 |
import gradio as gr
|
20 |
import glob
|
21 |
from inference_sam import segmentation_sam
|
22 |
-
|
|
|
23 |
import pathlib
|
|
|
|
|
24 |
|
25 |
if not os.path.exists('images'):
|
26 |
REPO_ID='Serrelab/image_examples_gradio'
|
27 |
snapshot_download(repo_id=REPO_ID, token=os.environ.get('READ_TOKEN'),repo_type='dataset',local_dir='images')
|
28 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
29 |
|
30 |
def segment_image(input_image):
|
31 |
img = segmentation_sam(input_image)
|
@@ -34,24 +60,54 @@ def segment_image(input_image):
|
|
34 |
def classify_image(input_image, model_name):
|
35 |
if 'Rock 170' ==model_name:
|
36 |
from inference_resnet import inference_resnet_finer
|
37 |
-
|
|
|
38 |
return result
|
39 |
elif 'Mummified 170' ==model_name:
|
40 |
from inference_resnet import inference_resnet_finer
|
41 |
-
|
|
|
42 |
return result
|
43 |
if 'Fossils 19' ==model_name:
|
44 |
from inference_beit import inference_dino
|
|
|
45 |
return inference_dino(input_image,model_name)
|
46 |
return None
|
47 |
|
48 |
-
def
|
49 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
50 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
51 |
|
52 |
with gr.Blocks(theme='sudeepshouche/minimalist') as demo:
|
53 |
|
54 |
-
with gr.Tab("
|
55 |
|
56 |
with gr.Row():
|
57 |
with gr.Column():
|
@@ -64,10 +120,10 @@ with gr.Blocks(theme='sudeepshouche/minimalist') as demo:
|
|
64 |
#classify_segmented_button = gr.Button("Classify Segmented Image")
|
65 |
|
66 |
with gr.Column():
|
67 |
-
|
68 |
-
["Mummified 170", "Rock 170"
|
69 |
multiselect=False,
|
70 |
-
value=
|
71 |
label="Model",
|
72 |
interactive=True,
|
73 |
)
|
@@ -81,24 +137,24 @@ with gr.Blocks(theme='sudeepshouche/minimalist') as demo:
|
|
81 |
samples=[[path.as_posix()] for path in paths if 'leaves' in str(path) ][:19]
|
82 |
examples_leaves = gr.Examples(samples, inputs=input_image,examples_per_page=5,label='Leaves Examples from the dataset')
|
83 |
|
84 |
-
with gr.Accordion("Using Diffuser"):
|
85 |
-
|
86 |
-
|
87 |
-
|
88 |
-
|
89 |
-
|
90 |
-
|
91 |
-
|
92 |
|
93 |
|
94 |
with gr.Accordion("Explanations "):
|
95 |
gr.Markdown("Computing Explanations from the model")
|
96 |
with gr.Row():
|
97 |
-
original_input = gr.Image(label="Original Frame")
|
98 |
saliency = gr.Image(label="saliency")
|
99 |
-
gradcam = gr.Image(label='
|
100 |
-
guided_gradcam = gr.Image(label='
|
101 |
-
guided_backprop = gr.Image(label='guided backprop')
|
102 |
generate_explanations = gr.Button("Generate Explanations")
|
103 |
|
104 |
with gr.Accordion('Closest Images'):
|
@@ -112,8 +168,10 @@ with gr.Blocks(theme='sudeepshouche/minimalist') as demo:
|
|
112 |
find_closest_btn = gr.Button("Find Closest Images")
|
113 |
|
114 |
segment_button.click(segment_image, inputs=input_image, outputs=segmented_image)
|
115 |
-
classify_image_button.click(classify_image, inputs=[input_image,
|
116 |
-
|
|
|
|
|
117 |
|
118 |
demo.queue()
|
119 |
|
|
|
9 |
subprocess.call('pip install git+https://github.com/cocodataset/panopticapi.git'.split())
|
10 |
subprocess.call('pip install python-dotenv'.split())
|
11 |
subprocess.call('pip install torch torchvision '.split())
|
12 |
+
subprocess.call('pip install xplique'.split())
|
13 |
|
14 |
import gradio as gr
|
15 |
from huggingface_hub import snapshot_download
|
|
|
20 |
import gradio as gr
|
21 |
import glob
|
22 |
from inference_sam import segmentation_sam
|
23 |
+
from explanations import explain
|
24 |
+
from inference_resnet import get_triplet_model
|
25 |
import pathlib
|
26 |
+
import tensorflow as tf
|
27 |
+
from closest_sample import get_images
|
28 |
|
29 |
if not os.path.exists('images'):
|
30 |
REPO_ID='Serrelab/image_examples_gradio'
|
31 |
snapshot_download(repo_id=REPO_ID, token=os.environ.get('READ_TOKEN'),repo_type='dataset',local_dir='images')
|
32 |
|
33 |
+
def get_model(model_name):
|
34 |
+
|
35 |
+
|
36 |
+
if model_name=='Mummified 170':
|
37 |
+
n_classes = 170
|
38 |
+
model = get_triplet_model(input_shape = (600, 600, 3),
|
39 |
+
embedding_units = 256,
|
40 |
+
embedding_depth = 2,
|
41 |
+
backbone_class=tf.keras.applications.ResNet50V2,
|
42 |
+
nb_classes = n_classes,load_weights=False,finer_model=True,backbone_name ='Resnet50v2')
|
43 |
+
model.load_weights('model_classification/mummified-170.h5')
|
44 |
+
elif model_name=='Rock 170':
|
45 |
+
n_classes = 171
|
46 |
+
model = get_triplet_model(input_shape = (600, 600, 3),
|
47 |
+
embedding_units = 256,
|
48 |
+
embedding_depth = 2,
|
49 |
+
backbone_class=tf.keras.applications.ResNet50V2,
|
50 |
+
nb_classes = n_classes,load_weights=False,finer_model=True,backbone_name ='Resnet50v2')
|
51 |
+
model.load_weights('model_classification/rock-170.h5')
|
52 |
+
else:
|
53 |
+
return 'Error'
|
54 |
+
return model,n_classes
|
55 |
|
56 |
def segment_image(input_image):
|
57 |
img = segmentation_sam(input_image)
|
|
|
60 |
def classify_image(input_image, model_name):
|
61 |
if 'Rock 170' ==model_name:
|
62 |
from inference_resnet import inference_resnet_finer
|
63 |
+
model,n_classes= get_model(model_name)
|
64 |
+
result = inference_resnet_finer(input_image,model,size=600,n_classes=n_classes)
|
65 |
return result
|
66 |
elif 'Mummified 170' ==model_name:
|
67 |
from inference_resnet import inference_resnet_finer
|
68 |
+
model, n_classes= get_model(model_name)
|
69 |
+
result = inference_resnet_finer(input_image,model,size=600,n_classes=n_classes)
|
70 |
return result
|
71 |
if 'Fossils 19' ==model_name:
|
72 |
from inference_beit import inference_dino
|
73 |
+
model,n_classes = get_model(model_name)
|
74 |
return inference_dino(input_image,model_name)
|
75 |
return None
|
76 |
|
77 |
+
def get_embeddings(input_image,model_name):
|
78 |
+
if 'Rock 170' ==model_name:
|
79 |
+
from inference_resnet import inference_resnet_embedding
|
80 |
+
model,n_classes= get_model(model_name)
|
81 |
+
result = inference_resnet_embedding(input_image,model,size=600,n_classes=n_classes)
|
82 |
+
return result
|
83 |
+
elif 'Mummified 170' ==model_name:
|
84 |
+
from inference_resnet import inference_resnet_embedding
|
85 |
+
model, n_classes= get_model(model_name)
|
86 |
+
result = inference_resnet_embedding(input_image,model,size=600,n_classes=n_classes)
|
87 |
+
return result
|
88 |
+
if 'Fossils 19' ==model_name:
|
89 |
+
from inference_beit import inference_dino
|
90 |
+
model,n_classes = get_model(model_name)
|
91 |
+
return inference_dino(input_image,model_name)
|
92 |
+
return None
|
93 |
+
|
94 |
|
95 |
+
def find_closest(input_image,model_name):
|
96 |
+
embedding = get_embeddings(input_image,model_name)
|
97 |
+
paths = get_images(embedding)
|
98 |
+
return paths
|
99 |
+
|
100 |
+
def explain_image(input_image,model_name):
|
101 |
+
model,n_classes= get_model(model_name)
|
102 |
+
saliency, integrated, smoothgrad = explain(model,input_image,n_classes=n_classes)
|
103 |
+
#original = saliency + integrated + smoothgrad
|
104 |
+
print('done')
|
105 |
+
return saliency, integrated, smoothgrad,
|
106 |
+
|
107 |
|
108 |
with gr.Blocks(theme='sudeepshouche/minimalist') as demo:
|
109 |
|
110 |
+
with gr.Tab(" Florrissant Fossils"):
|
111 |
|
112 |
with gr.Row():
|
113 |
with gr.Column():
|
|
|
120 |
#classify_segmented_button = gr.Button("Classify Segmented Image")
|
121 |
|
122 |
with gr.Column():
|
123 |
+
model_name = gr.Dropdown(
|
124 |
+
["Mummified 170", "Rock 170"],
|
125 |
multiselect=False,
|
126 |
+
value="Rock 170",
|
127 |
label="Model",
|
128 |
interactive=True,
|
129 |
)
|
|
|
137 |
samples=[[path.as_posix()] for path in paths if 'leaves' in str(path) ][:19]
|
138 |
examples_leaves = gr.Examples(samples, inputs=input_image,examples_per_page=5,label='Leaves Examples from the dataset')
|
139 |
|
140 |
+
# with gr.Accordion("Using Diffuser"):
|
141 |
+
# with gr.Column():
|
142 |
+
# prompt = gr.Textbox(lines=1, label="Prompt")
|
143 |
+
# output_image = gr.Image(label="Output")
|
144 |
+
# generate_button = gr.Button("Generate Leave")
|
145 |
+
# with gr.Column():
|
146 |
+
# class_predicted2 = gr.Label(label='Class Predicted from diffuser')
|
147 |
+
# classify_button = gr.Button("Classify Image")
|
148 |
|
149 |
|
150 |
with gr.Accordion("Explanations "):
|
151 |
gr.Markdown("Computing Explanations from the model")
|
152 |
with gr.Row():
|
153 |
+
#original_input = gr.Image(label="Original Frame")
|
154 |
saliency = gr.Image(label="saliency")
|
155 |
+
gradcam = gr.Image(label='integraged gradients')
|
156 |
+
guided_gradcam = gr.Image(label='gradcam')
|
157 |
+
#guided_backprop = gr.Image(label='guided backprop')
|
158 |
generate_explanations = gr.Button("Generate Explanations")
|
159 |
|
160 |
with gr.Accordion('Closest Images'):
|
|
|
168 |
find_closest_btn = gr.Button("Find Closest Images")
|
169 |
|
170 |
segment_button.click(segment_image, inputs=input_image, outputs=segmented_image)
|
171 |
+
classify_image_button.click(classify_image, inputs=[input_image,model_name], outputs=class_predicted)
|
172 |
+
generate_explanations.click(explain_image, inputs=[input_image,model_name], outputs=[saliency,gradcam,guided_gradcam])
|
173 |
+
find_closest_btn.click(find_closest, inputs=[input_image,model_name], outputs=[closest_image_0,closest_image_1,closest_image_2,closest_image_3,closest_image_4])
|
174 |
+
#classify_segmented_button.click(classify_image, inputs=[segmented_image,model_name], outputs=class_predicted)
|
175 |
|
176 |
demo.queue()
|
177 |
|
inference_resnet.py
CHANGED
@@ -12,9 +12,10 @@ from huggingface_hub import snapshot_download
|
|
12 |
from labels import lookup_170
|
13 |
import numpy as np
|
14 |
|
|
|
15 |
|
16 |
-
REPO_ID='Serrelab/fossil_classification_models'
|
17 |
-
snapshot_download(repo_id=REPO_ID, token=os.environ.get('READ_TOKEN'),repo_type='model',local_dir='model_classification')
|
18 |
|
19 |
|
20 |
def get_model(base_arch='Nasnet',weights='imagenet',input_shape=(600,600,3),classes=64500):
|
@@ -146,19 +147,19 @@ def parse_results(top_n,logits):
|
|
146 |
results[label] = float(logits[n])
|
147 |
return results
|
148 |
|
149 |
-
def
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
150 |
|
151 |
-
model = get_triplet_model(input_shape = (size, size, 3),
|
152 |
-
embedding_units = 256,
|
153 |
-
embedding_depth = 2,
|
154 |
-
backbone_class=tf.keras.applications.ResNet50V2,
|
155 |
-
nb_classes = n_classes,load_weights=False,finer_model=True,backbone_name ='Resnet50v2')
|
156 |
-
if type_model=='Mummified 170':
|
157 |
-
model.load_weights('model_classification/mummified-170.h5')
|
158 |
-
elif type_model=='Rock 170':
|
159 |
-
model.load_weights('model_classification/rock-170.h5')
|
160 |
-
else:
|
161 |
-
return 'Error'
|
162 |
cropped = _clever_crop(x,(size,size))[0]
|
163 |
prep = preprocess(cropped,size=size)
|
164 |
logits = tf.nn.softmax(model.predict(np.array([prep]))[1][0]).cpu().numpy()
|
|
|
12 |
from labels import lookup_170
|
13 |
import numpy as np
|
14 |
|
15 |
+
if not os.path.exists('model_classification'):
|
16 |
|
17 |
+
REPO_ID='Serrelab/fossil_classification_models'
|
18 |
+
snapshot_download(repo_id=REPO_ID, token=os.environ.get('READ_TOKEN'),repo_type='model',local_dir='model_classification')
|
19 |
|
20 |
|
21 |
def get_model(base_arch='Nasnet',weights='imagenet',input_shape=(600,600,3),classes=64500):
|
|
|
147 |
results[label] = float(logits[n])
|
148 |
return results
|
149 |
|
150 |
+
def inference_resnet_embedding(x,model,size=576,n_classes=170,n_top=10):
|
151 |
+
|
152 |
+
|
153 |
+
cropped = _clever_crop(x,(size,size))[0]
|
154 |
+
prep = preprocess(cropped,size=size)
|
155 |
+
embedding = model.predict(np.array([prep]))[0][0]
|
156 |
+
|
157 |
+
|
158 |
+
return embedding
|
159 |
+
|
160 |
+
def inference_resnet_finer(x,model,size=576,n_classes=170,n_top=10):
|
161 |
+
|
162 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
163 |
cropped = _clever_crop(x,(size,size))[0]
|
164 |
prep = preprocess(cropped,size=size)
|
165 |
logits = tf.nn.softmax(model.predict(np.array([prep]))[1][0]).cpu().numpy()
|
inference_sam.py
CHANGED
@@ -12,10 +12,11 @@ from math import ceil
|
|
12 |
import os
|
13 |
from huggingface_hub import snapshot_download
|
14 |
|
15 |
-
|
16 |
-
|
|
|
17 |
|
18 |
-
sam = sam_model_registry["default"]("model/sam_02-06_dice_mse_0.pth")
|
19 |
sam.cuda()
|
20 |
predictor = SamPredictor(sam)
|
21 |
|
@@ -172,4 +173,4 @@ def segmentation_sam(x,SIZE=384):
|
|
172 |
data = np.frombuffer(fig.canvas.tostring_rgb(), dtype=np.uint8)
|
173 |
data = data.reshape(fig.canvas.get_width_height()[::-1] + (3,))
|
174 |
plt.close()
|
175 |
-
return data
|
|
|
12 |
import os
|
13 |
from huggingface_hub import snapshot_download
|
14 |
|
15 |
+
if not os.path.exists('model'):
|
16 |
+
REPO_ID='Serrelab/SAM_Leaves'
|
17 |
+
snapshot_download(repo_id=REPO_ID, token=os.environ.get('READ_TOKEN'),repo_type='model',local_dir='model')
|
18 |
|
19 |
+
sam = sam_model_registry["default"]("/home/irodri15/Documents/Projects/Fossils/fossil_app/model/sam_02-06_dice_mse_0.pth")
|
20 |
sam.cuda()
|
21 |
predictor = SamPredictor(sam)
|
22 |
|
|
|
173 |
data = np.frombuffer(fig.canvas.tostring_rgb(), dtype=np.uint8)
|
174 |
data = data.reshape(fig.canvas.get_width_height()[::-1] + (3,))
|
175 |
plt.close()
|
176 |
+
return data
|
pre-requirements.txt
CHANGED
@@ -3,4 +3,5 @@ opencv-python-headless==4.5.5.64
|
|
3 |
openmim==0.1.5
|
4 |
torch==1.11.0
|
5 |
torchvision==0.12.0
|
6 |
-
tensorflow==2.8
|
|
|
|
3 |
openmim==0.1.5
|
4 |
torch==1.11.0
|
5 |
torchvision==0.12.0
|
6 |
+
tensorflow==2.8
|
7 |
+
xplique
|