RabbitDaisuke commited on
Commit
14fdcae
·
verified ·
1 Parent(s): f0d9044

Update FP32 ONNX from 5d7918b55b92

Browse files
Files changed (4) hide show
  1. LICENSE +21 -0
  2. README.md +23 -0
  3. zipdepth-base-768-fp32.json +104 -0
  4. zipdepth-base-768-fp32.onnx +3 -0
LICENSE ADDED
@@ -0,0 +1,21 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ MIT License
2
+
3
+ Copyright (c) 2026 Fabio Tosi
4
+
5
+ Permission is hereby granted, free of charge, to any person obtaining a copy
6
+ of this software and associated documentation files (the "Software"), to deal
7
+ in the Software without restriction, including without limitation the rights
8
+ to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
9
+ copies of the Software, and to permit persons to whom the Software is
10
+ furnished to do so, subject to the following conditions:
11
+
12
+ The above copyright notice and this permission notice shall be included in all
13
+ copies or substantial portions of the Software.
14
+
15
+ THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
16
+ IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
17
+ FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
18
+ AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
19
+ LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
20
+ OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
21
+ SOFTWARE.
README.md ADDED
@@ -0,0 +1,23 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ license: mit
3
+ library_name: onnx
4
+ pipeline_tag: depth-estimation
5
+ tags:
6
+ - onnx
7
+ - webgpu
8
+ - zipdepth
9
+ ---
10
+ # ZipDepth ONNX
11
+ FP32 ONNX conversion of [Fabio Tosi et al.'s ZipDepth](https://github.com/fabiotosi92/ZipDepth), with the original learned attention and convex upsampling preserved. No quantization or retraining is applied.
12
+ ## License and provenance
13
+ The upstream repository distributes its code and checkpoints under its MIT LICENSE; the original copyright and permission notice is included in LICENSE. This conversion is not an official upstream release. ZipDepth was distilled from Depth Anything V2 Large; this card does not relicense the teacher or its original datasets, which are not included here.
14
+ ## Input and output
15
+ Input `image` is FP32 RGB divided by 255, NCHW `[1,3,768,768]`. Preserve the source aspect ratio when resizing and pad the remaining area with black. Ignore padded regions when displaying or exporting scene points.
16
+ Output `inverse_depth` is FP32 `[1,1,768,768]`, affine-invariant inverse depth: larger means nearer. It is not a distance in meters. Calibrated distance requires a scale and shift in inverse-depth space. A relative preview may use `z = scale / (raw + shift)` with positive denominators, while retaining the raw output unchanged.
17
+ ## Validation
18
+ The JSON manifest contains checkpoint and ONNX SHA-256 hashes and numerical comparisons with the original PyTorch model on a public image, RGB noise and black input. Browser latency and WebGPU operator support must be measured separately; CPU conversion checks are not browser benchmarks.
19
+ ## Build
20
+ Source: https://github.com/DaisukeDaisuke/ZipDepth
21
+ Every successful build updates this same repository. Use a Hugging Face commit revision to pin an immutable model download.
22
+ Source commit: `5d7918b55b9204ee24fc76f0c16778d612402d3e`
23
+ ONNX SHA-256: `7926169714cc7e310408f4640f650db3253be1453ad185384bddbea13339795b`
zipdepth-base-768-fp32.json ADDED
@@ -0,0 +1,104 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "model": "ZipDepth-base",
3
+ "precision": "float32",
4
+ "output_semantics": "affine-invariant inverse depth; larger means nearer",
5
+ "preprocess": "RGB / 255, NCHW; aspect-preserving resize and black padding",
6
+ "checkpoint_sha256": "a55910bb0b99c8c5e641cb9206e810b269690ad94e8a2ef08c827c4679391a65",
7
+ "source_commit": "5d7918b55b9204ee24fc76f0c16778d612402d3e",
8
+ "filename": "zipdepth-base-768-fp32.onnx",
9
+ "sha256": "7926169714cc7e310408f4640f650db3253be1453ad185384bddbea13339795b",
10
+ "bytes": 24624186,
11
+ "input_name": "image",
12
+ "output_name": "inverse_depth",
13
+ "input_shape": [
14
+ 1,
15
+ 3,
16
+ 768,
17
+ 768
18
+ ],
19
+ "output_shape": [
20
+ 1,
21
+ 1,
22
+ 768,
23
+ 768
24
+ ],
25
+ "opset": 17,
26
+ "operators": [
27
+ "Add",
28
+ "AveragePool",
29
+ "BatchNormalization",
30
+ "Cast",
31
+ "Concat",
32
+ "Constant",
33
+ "ConstantOfShape",
34
+ "Conv",
35
+ "DepthToSpace",
36
+ "Div",
37
+ "Gather",
38
+ "GlobalAveragePool",
39
+ "MatMul",
40
+ "MaxPool",
41
+ "Mul",
42
+ "Pad",
43
+ "Range",
44
+ "ReduceMean",
45
+ "ReduceSum",
46
+ "Relu",
47
+ "Reshape",
48
+ "Resize",
49
+ "Shape",
50
+ "Sigmoid",
51
+ "Slice",
52
+ "Softmax",
53
+ "Sub",
54
+ "Transpose",
55
+ "Unsqueeze"
56
+ ],
57
+ "validation": [
58
+ {
59
+ "label": "fusion/public-image",
60
+ "max_absolute_error": 5.029141902923584e-07,
61
+ "mean_absolute_error": 1.8642849042294074e-08,
62
+ "output_min": 0.0,
63
+ "output_max": 0.17091958224773407
64
+ },
65
+ {
66
+ "label": "fusion/random-rgb",
67
+ "max_absolute_error": 1.043081283569336e-07,
68
+ "mean_absolute_error": 1.2551014449968534e-08,
69
+ "output_min": 0.009747229516506195,
70
+ "output_max": 0.09417390078306198
71
+ },
72
+ {
73
+ "label": "fusion/black",
74
+ "max_absolute_error": 3.650784492492676e-07,
75
+ "mean_absolute_error": 9.113544763295067e-08,
76
+ "output_min": 0.04116407036781311,
77
+ "output_max": 0.11994780600070953
78
+ },
79
+ {
80
+ "label": "onnx/public-image",
81
+ "max_absolute_error": 6.034970283508301e-07,
82
+ "mean_absolute_error": 5.203410324838842e-08,
83
+ "output_min": 0.0,
84
+ "output_max": 0.17091967165470123
85
+ },
86
+ {
87
+ "label": "onnx/random-rgb",
88
+ "max_absolute_error": 1.564621925354004e-07,
89
+ "mean_absolute_error": 2.4493134631597968e-08,
90
+ "output_min": 0.009747234173119068,
91
+ "output_max": 0.09417388588190079
92
+ },
93
+ {
94
+ "label": "onnx/black",
95
+ "max_absolute_error": 1.2293457984924316e-06,
96
+ "mean_absolute_error": 3.444413323450135e-07,
97
+ "output_min": 0.041164495050907135,
98
+ "output_max": 0.11994778364896774
99
+ }
100
+ ],
101
+ "checkpoint_unused_keys": [],
102
+ "torch_version": "2.5.1+cpu",
103
+ "onnxruntime_version": "1.20.1"
104
+ }
zipdepth-base-768-fp32.onnx ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:7926169714cc7e310408f4640f650db3253be1453ad185384bddbea13339795b
3
+ size 24624186