Hugging Face's logo

zhyever
/
patchfusion_depth_anything_vitl14

Transformers
PyTorch
Inference Endpoints
Model card Files Community
zhyever commited on
Commit
61e31a6
•
1 Parent(s): db174cd

Upload checkpoints

Browse files
Files changed (2) hide show
  1. config.json +171 -0
  2. pytorch_model.bin +3 -0
config.json ADDED
@@ -0,0 +1,171 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "coarse_branch": {
3
+ "attractor_alpha": 1000,
4
+ "attractor_gamma": 2,
5
+ "attractor_kind": "mean",
6
+ "attractor_type": "inv",
7
+ "aug": true,
8
+ "bin_centers_type": "softplus",
9
+ "bin_embedding_dim": 128,
10
+ "clip_grad": 0.1,
11
+ "dataset": "nyu",
12
+ "depth_anything": true,
13
+ "distributed": true,
14
+ "do_resize": false,
15
+ "force_keep_ar": true,
16
+ "freeze_midas_bn": true,
17
+ "gpu": "NULL",
18
+ "img_size": [
19
+ 392,
20
+ 518
21
+ ],
22
+ "inverse_midas": false,
23
+ "log_images_every": 0.1,
24
+ "max_depth": 80,
25
+ "max_temp": 50.0,
26
+ "max_translation": 100,
27
+ "memory_efficient": true,
28
+ "midas_model_type": "vitl",
29
+ "min_depth": 0.001,
30
+ "min_temp": 0.0212,
31
+ "model": "zoedepth",
32
+ "n_attractors": [
33
+ 16,
34
+ 8,
35
+ 4,
36
+ 1
37
+ ],
38
+ "n_bins": 64,
39
+ "name": "ZoeDepth",
40
+ "notes": "",
41
+ "output_distribution": "logbinomial",
42
+ "prefetch": false,
43
+ "pretrained_resource": "local::./work_dir/DepthAnything_vitl.pt",
44
+ "print_losses": false,
45
+ "project": "ZoeDepth",
46
+ "random_crop": false,
47
+ "random_translate": false,
48
+ "root": ".",
49
+ "save_dir": "",
50
+ "shared_dict": "NULL",
51
+ "tags": "",
52
+ "train_midas": true,
53
+ "translate_prob": 0.2,
54
+ "type": "DA-ZoeDepth",
55
+ "uid": "NULL",
56
+ "use_amp": false,
57
+ "use_pretrained_midas": true,
58
+ "use_shared_dict": false,
59
+ "validate_every": 0.25,
60
+ "version_name": "v1",
61
+ "workers": 16
62
+ },
63
+ "fine_branch": {
64
+ "attractor_alpha": 1000,
65
+ "attractor_gamma": 2,
66
+ "attractor_kind": "mean",
67
+ "attractor_type": "inv",
68
+ "aug": true,
69
+ "bin_centers_type": "softplus",
70
+ "bin_embedding_dim": 128,
71
+ "clip_grad": 0.1,
72
+ "dataset": "nyu",
73
+ "depth_anything": true,
74
+ "distributed": true,
75
+ "do_resize": false,
76
+ "force_keep_ar": true,
77
+ "freeze_midas_bn": true,
78
+ "gpu": "NULL",
79
+ "img_size": [
80
+ 392,
81
+ 518
82
+ ],
83
+ "inverse_midas": false,
84
+ "log_images_every": 0.1,
85
+ "max_depth": 80,
86
+ "max_temp": 50.0,
87
+ "max_translation": 100,
88
+ "memory_efficient": true,
89
+ "midas_model_type": "vitl",
90
+ "min_depth": 0.001,
91
+ "min_temp": 0.0212,
92
+ "model": "zoedepth",
93
+ "n_attractors": [
94
+ 16,
95
+ 8,
96
+ 4,
97
+ 1
98
+ ],
99
+ "n_bins": 64,
100
+ "name": "ZoeDepth",
101
+ "notes": "",
102
+ "output_distribution": "logbinomial",
103
+ "prefetch": false,
104
+ "pretrained_resource": "local::./work_dir/DepthAnything_vitl.pt",
105
+ "print_losses": false,
106
+ "project": "ZoeDepth",
107
+ "random_crop": false,
108
+ "random_translate": false,
109
+ "root": ".",
110
+ "save_dir": "",
111
+ "shared_dict": "NULL",
112
+ "tags": "",
113
+ "train_midas": true,
114
+ "translate_prob": 0.2,
115
+ "type": "DA-ZoeDepth",
116
+ "uid": "NULL",
117
+ "use_amp": false,
118
+ "use_pretrained_midas": true,
119
+ "use_shared_dict": false,
120
+ "validate_every": 0.25,
121
+ "version_name": "v1",
122
+ "workers": 16
123
+ },
124
+ "guided_fusion": {
125
+ "g2l": true,
126
+ "in_channels": [
127
+ 32,
128
+ 256,
129
+ 256,
130
+ 256,
131
+ 256,
132
+ 256
133
+ ],
134
+ "n_channels": 5,
135
+ "num_patches": [
136
+ 203056,
137
+ 66304,
138
+ 16576,
139
+ 4144,
140
+ 1036,
141
+ 266
142
+ ],
143
+ "patch_process_shape": [
144
+ 392,
145
+ 518
146
+ ],
147
+ "type": "GuidedFusionPatchFusion"
148
+ },
149
+ "image_raw_shape": [
150
+ 2160,
151
+ 3840
152
+ ],
153
+ "max_depth": 80,
154
+ "min_depth": 0.001,
155
+ "patch_process_shape": [
156
+ 392,
157
+ 518
158
+ ],
159
+ "patch_split_num": [
160
+ 4,
161
+ 4
162
+ ],
163
+ "pretrain_model": [
164
+ "./work_dir/depthanything_vitl_u4k/coarse_pretrain/checkpoint_24.pth",
165
+ "./work_dir/depthanything_vitl_u4k/fine_pretrain/checkpoint_24.pth"
166
+ ],
167
+ "sigloss": {
168
+ "type": "SILogLoss"
169
+ },
170
+ "transformers_version": "4.36.2"
171
+ }
pytorch_model.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:756398ef543adbd63cf4bf24bdcbc61980dc8f7b25d346e14f663d46ef45f752
3
+ size 3064827728