Hugging Face's logo

deepang
/
EAT

Model card Files
xet
Community
lealaxy commited on
Commit
fda0535
·
1 Parent(s): 568f346

update

Browse files
Files changed (4) hide show
  1. Kvasir/log/EAT-Kvasir-SEG_val_log.txt +266 -0
  2. Kvasir/weight/pytorch_model.bin +3 -0
  3. POLYPGEN/logs/EAT_PolypGEN_log.txt +0 -0
  4. POLYPGEN/weight/pytorch_model.bin +3 -0
Kvasir/log/EAT-Kvasir-SEG_val_log.txt ADDED
@@ -0,0 +1,266 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ <EasyDict 0x79ac5b1e4900
2
+ 'dataset': <EasyDict 0x79ab5c0d6430
3
+ 'CVC_ClinicDB': <EasyDict 0x79ab52c53060
4
+ 'batch_size': 8,
5
+ 'data_root': '/mnt/d/dataset/CVC-ClinicDB/',
6
+ 'image_size': 352,
7
+ 'num_workers': 4,
8
+ 'train_ratio': 0.8
9
+ >,
10
+ 'EDD_seg': <EasyDict 0x79ab52c52250
11
+ 'batch_size': 8,
12
+ 'data_root': '/dataset/cv/seg/EDD2020/',
13
+ 'image_mean': 0.5,
14
+ 'image_size': 352,
15
+ 'image_std': 0.5,
16
+ 'num_workers': 4,
17
+ 'train_ratio': 0.8
18
+ >,
19
+ 'Kvasir_SEG': <EasyDict 0x79ab52c521b0
20
+ 'batch_size': 8,
21
+ 'data_root': '/mnt/d/dataset/Kvasir-SEG/',
22
+ 'image_size': 352,
23
+ 'num_workers': 4,
24
+ 'train_ratio': 0.8
25
+ >,
26
+ 'PolypGen': <EasyDict 0x79ab52c52200
27
+ 'batch_size': 8,
28
+ 'data_root': '/mnt/d/dataset/PolypGen/',
29
+ 'image_mean': 0.5,
30
+ 'image_size': 352,
31
+ 'image_std': 0.5,
32
+ 'num_workers': 4,
33
+ 'train_ratio': 0.8
34
+ >,
35
+ 'Sun_seg': <EasyDict 0x79ab52c52110
36
+ 'batch_size': 8,
37
+ 'data_root': '/mnt/d/dataset/SUN-SEG/',
38
+ 'image_mean': 0.5,
39
+ 'image_size': 352,
40
+ 'image_std': 0.5,
41
+ 'num_workers': 4,
42
+ 'train_ratio': 0.8
43
+ >
44
+ >,
45
+ 'finetune': <EasyDict 0x79ab52fd9620
46
+ 'checkpoint': 'EAT_kva_best',
47
+ 'model_choose': 'EAT'
48
+ >,
49
+ 'models': <EasyDict 0x79ab52c522a0
50
+ 'FCBFormer': <EasyDict 0x79ab52c52610
51
+ 'branch1': <EasyDict 0x79ab52c52750
52
+ 'model_dir': '/home/yoikl/.cache/huggingface/forget/pvt_v2_b3.pth',
53
+ 'num_class': 1,
54
+ 'size': 352
55
+ >,
56
+ 'branch5': <EasyDict 0x79ab52c52840
57
+ 'model_dir': '/home/yoikl/.cache/huggingface/forget/pvt_v2_b3.pth',
58
+ 'num_class': 5,
59
+ 'size': 352
60
+ >
61
+ >,
62
+ 'LDNet': <EasyDict 0x79ab52c52a70
63
+ 'branch1': <EasyDict 0x79ab52c52c00
64
+ 'conv_kernel_size': 1,
65
+ 'num_classes': 1,
66
+ 'unified_channels': 64
67
+ >,
68
+ 'branch5': <EasyDict 0x79ab52c52cf0
69
+ 'conv_kernel_size': 1,
70
+ 'num_classes': 5,
71
+ 'unified_channels': 64
72
+ >
73
+ >,
74
+ 'cfp_net': <EasyDict 0x79ab52c529d0
75
+ 'branch1': <EasyDict 0x79ab52c52b10
76
+ 'block_1': 2,
77
+ 'block_2': 6,
78
+ 'classes': 1
79
+ >,
80
+ 'branch5': <EasyDict 0x79ab52c52c50
81
+ 'block_1': 2,
82
+ 'block_2': 6,
83
+ 'classes': 5
84
+ >
85
+ >,
86
+ 'cvc_unetr': <EasyDict 0x79ab52c522f0
87
+ 'branch1': <EasyDict 0x79ab52c52430
88
+ 'dims': [64, 128, 320, 512],
89
+ 'in_channels': 3,
90
+ 'kernel_size': 3,
91
+ 'mlp_ratio': 4,
92
+ 'model_dir': '/home/yoikl/.cache/huggingface/forget/pvt_v2_b2.pth',
93
+ 'out_channels': 1,
94
+ 'out_dim': 32
95
+ >,
96
+ 'branch5': <EasyDict 0x79ab52c52520
97
+ 'dims': [64, 128, 320, 512],
98
+ 'in_channels': 3,
99
+ 'kernel_size': 3,
100
+ 'mlp_ratio': 4,
101
+ 'model_dir': '/home/yoikl/.cache/huggingface/forget/pvt_v2_b2.pth',
102
+ 'out_channels': 5,
103
+ 'out_dim': 32
104
+ >
105
+ >,
106
+ 'EAT': <EasyDict 0x79ab52c52480
107
+ 'branch1': <EasyDict 0x79ab52c523e0
108
+ 'L_feature_os': 1.0,
109
+ 'L_feature_v4': True,
110
+ 'dims': [64, 128, 320, 512],
111
+ 'gla_os': [3.0, 1.0],
112
+ 'gla_v4': [True, False],
113
+ 'in_channels': 3,
114
+ 'kernel_size': 3,
115
+ 'mlp_ratio': 4,
116
+ 'model_dir': '/home/yoikl/.cache/huggingface/forget/pvt_v2_b2.pth',
117
+ 'out_channels': 1,
118
+ 'out_dim': 32
119
+ >,
120
+ 'branch5': <EasyDict 0x79ab53061c60
121
+ 'dims': [64, 128, 320, 512],
122
+ 'in_channels': 3,
123
+ 'kernel_size': 3,
124
+ 'mlp_ratio': 4,
125
+ 'model_dir': '/home/yoikl/.cache/huggingface/forget/pvt_v2_b2.pth',
126
+ 'out_channels': 5,
127
+ 'out_dim': 32
128
+ >,
129
+ 'visualize': <EasyDict 0x79ab52c524d0
130
+ 'cam_target_layer': 'fuse2'
131
+ >
132
+ >,
133
+ 'duat': <EasyDict 0x79ab52c525c0
134
+ 'branch1': <EasyDict 0x79ab52c52700
135
+ 'dim': 32,
136
+ 'dims': [64, 128, 320, 512],
137
+ 'in_channels': 3,
138
+ 'model_dir': '/home/yoikl/.cache/huggingface/forget/pvt_v2_b2.pth',
139
+ 'out_channels': 1
140
+ >,
141
+ 'branch5': <EasyDict 0x79ab52c52570
142
+ 'dim': 32,
143
+ 'dims': [64, 128, 320, 512],
144
+ 'in_channels': 3,
145
+ 'model_dir': '/home/yoikl/.cache/huggingface/forget/pvt_v2_b2.pth',
146
+ 'out_channels': 5
147
+ >
148
+ >,
149
+ 'swin_unetr': <EasyDict 0x79ab52c526b0
150
+ 'branch1': <EasyDict 0x79ab52c527f0
151
+ 'img_size': [352, 352],
152
+ 'in_channels': 3,
153
+ 'out_channels': 1,
154
+ 'spatial_dims': 2,
155
+ 'use_checkpoint': True
156
+ >,
157
+ 'branch5': <EasyDict 0x79ab52c528e0
158
+ 'img_size': [352, 352],
159
+ 'in_channels': 3,
160
+ 'out_channels': 5,
161
+ 'spatial_dims': 2,
162
+ 'use_checkpoint': True
163
+ >
164
+ >,
165
+ 'trans_unet': <EasyDict 0x79ab52c52a20
166
+ 'branch1': <EasyDict 0x79ab52c52bb0
167
+ 'block_num': 8,
168
+ 'class_num': 1,
169
+ 'head_num': 4,
170
+ 'img_dim': 352,
171
+ 'in_channels': 3,
172
+ 'mlp_dim': 512,
173
+ 'out_channels': 128,
174
+ 'patch_dim': 16
175
+ >,
176
+ 'branch5': <EasyDict 0x79ab52c52ca0
177
+ 'block_num': 8,
178
+ 'class_num': 5,
179
+ 'head_num': 4,
180
+ 'img_dim': 352,
181
+ 'in_channels': 3,
182
+ 'mlp_dim': 512,
183
+ 'out_channels': 128,
184
+ 'patch_dim': 16
185
+ >
186
+ >,
187
+ 'u_netr': <EasyDict 0x79ab52c52980
188
+ 'branch1': <EasyDict 0x79ab52c52ac0
189
+ 'feature_size': 64,
190
+ 'img_size': 352,
191
+ 'in_channels': 3,
192
+ 'out_channels': 1,
193
+ 'spatial_dims': 2
194
+ >,
195
+ 'branch5': <EasyDict 0x79ab52c52b60
196
+ 'feature_size': 64,
197
+ 'img_size': 352,
198
+ 'in_channels': 3,
199
+ 'out_channels': 5,
200
+ 'spatial_dims': 2
201
+ >
202
+ >,
203
+ 'unet': <EasyDict 0x79ab52c52660
204
+ 'branch1': <EasyDict 0x79ab52c527a0
205
+ 'bilinear': False,
206
+ 'n_channels': 3,
207
+ 'n_classes': 1
208
+ >,
209
+ 'branch5': <EasyDict 0x79ab52c52890
210
+ 'bilinear': False,
211
+ 'n_channels': 3,
212
+ 'n_classes': 5
213
+ >
214
+ >
215
+ >,
216
+ 'shared path': '/home/yoikl/.cache/huggingface/forget/model_stores/',
217
+ 'trainer': <EasyDict 0x79ab52c2dd00
218
+ 'dataset_choose': 'Kvasir_SEG',
219
+ 'lr': 1e-05,
220
+ 'min_lr': 1e-07,
221
+ 'num_epochs': 4000,
222
+ 'optimizer': 'adamw',
223
+ 'pred_ratio_var': 0,
224
+ 'resume': True,
225
+ 'train_ratio': 0.8,
226
+ 'warmup': 3,
227
+ 'weight_decay': 0.05,
228
+ 'weight_decay_end': 0.04
229
+ >,
230
+ 'visualization': <EasyDict 0x79ab52c52340
231
+ 'img_path': './visualization/img/',
232
+ 'mask_path': './visualization/label/',
233
+ 'visualization_path': './visualization/output/'
234
+ >
235
+ >
236
+ Load Model...
237
+ Load Dataloader...
238
+ Successfully loaded the training model！
239
+ Validation [1/25] Loss: 1.65311 focal_loss 1.48653 dice_loss 0.16659
240
+ Validation [2/25] Loss: 0.28208 focal_loss 0.22234 dice_loss 0.05974
241
+ Validation [3/25] Loss: 1.19983 focal_loss 1.08626 dice_loss 0.11357
242
+ Validation [4/25] Loss: 1.85829 focal_loss 1.71635 dice_loss 0.14193
243
+ Validation [5/25] Loss: 0.29398 focal_loss 0.22048 dice_loss 0.07350
244
+ Validation [6/25] Loss: 0.34766 focal_loss 0.23794 dice_loss 0.10972
245
+ Validation [7/25] Loss: 0.41792 focal_loss 0.34445 dice_loss 0.07348
246
+ Validation [8/25] Loss: 0.50481 focal_loss 0.39103 dice_loss 0.11378
247
+ Validation [9/25] Loss: 0.54926 focal_loss 0.44895 dice_loss 0.10032
248
+ Validation [10/25] Loss: 1.45736 focal_loss 1.30696 dice_loss 0.15040
249
+ Validation [11/25] Loss: 0.39909 focal_loss 0.33849 dice_loss 0.06060
250
+ Validation [12/25] Loss: 1.42168 focal_loss 1.19001 dice_loss 0.23167
251
+ Validation [13/25] Loss: 1.25499 focal_loss 1.07108 dice_loss 0.18390
252
+ Validation [14/25] Loss: 2.02651 focal_loss 1.88316 dice_loss 0.14335
253
+ Validation [15/25] Loss: 0.36274 focal_loss 0.22355 dice_loss 0.13919
254
+ Validation [16/25] Loss: 0.13409 focal_loss 0.09684 dice_loss 0.03725
255
+ Validation [17/25] Loss: 0.51398 focal_loss 0.44316 dice_loss 0.07082
256
+ Validation [18/25] Loss: 0.77349 focal_loss 0.64868 dice_loss 0.12481
257
+ Validation [19/25] Loss: 0.51320 focal_loss 0.42167 dice_loss 0.09152
258
+ Validation [20/25] Loss: 0.66414 focal_loss 0.43682 dice_loss 0.22731
259
+ Validation [21/25] Loss: 1.58072 focal_loss 1.42365 dice_loss 0.15708
260
+ Validation [22/25] Loss: 1.05039 focal_loss 0.92069 dice_loss 0.12970
261
+ Validation [23/25] Loss: 2.00498 focal_loss 1.84157 dice_loss 0.16340
262
+ Validation [24/25] Loss: 3.16328 focal_loss 2.93746 dice_loss 0.22582
263
+ Validation [25/25] Loss: 0.95918 focal_loss 0.85015 dice_loss 0.10903
264
+ Epoch [1/4000] Validation metric {'Val/mean dice_metric': 0.9709388017654419, 'Val/mean miou_metric': 0.9785973429679871, 'Val/mean f1': 0.9638663530349731, 'Val/mean precision': 0.9675034880638123, 'Val/mean recall': 0.9602563977241516, 'Val/mean hd95_metric': 6.603846549987793}
265
+ Best acc: tensor([0.9709], device='cuda:0')
266
+ Best class : {'Val/mean dice_metric': 0.9709388017654419, 'Val/mean miou_metric': 0.9785973429679871, 'Val/mean f1': 0.9638663530349731, 'Val/mean precision': 0.9675034880638123, 'Val/mean recall': 0.9602563977241516, 'Val/mean hd95_metric': 6.603846549987793}
Kvasir/weight/pytorch_model.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:02ed23bdb1174080d3d0317d53c099294ad34931b3da5e83858ba1dbe3eb4782
3
+ size 101417089
POLYPGEN/logs/EAT_PolypGEN_log.txt ADDED
The diff for this file is too large to render. See raw diff
 
POLYPGEN/weight/pytorch_model.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:dcd5ecd1ec9c22e50c7212c09bfffbf9affe3fd3e5593dca91e564e77eded219
3
+ size 101417534