cy0307 commited on
Commit
b219d6a
·
verified ·
1 Parent(s): 5f60fa7

Upload from Ropedia Academy

Browse files
Files changed (4) hide show
  1. README.md +44 -3
  2. figure.png +2 -2
  3. metrics.json +11 -11
  4. sdf.pt +1 -1
README.md CHANGED
@@ -1,13 +1,54 @@
1
  ---
2
  license: mit
 
3
  tags:
4
  - ropedia-academy
5
  - educational
 
6
  ---
7
 
8
- # B_deepsdf_shape
9
 
10
- Trained from scratch in **Ropedia Academy** (https://chaoyue0307.github.io/ropedia-academy/) an educational lab. The checkpoint and full loss/eval history (`metrics.json`) are included.
11
 
 
12
 
13
- ![results](figure.png)
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
  ---
2
  license: mit
3
+ library_name: pytorch
4
  tags:
5
  - ropedia-academy
6
  - educational
7
+ - implicit-neural-representation
8
  ---
9
 
10
+ # Neural SDF (DeepSDF-style)
11
 
12
+ An MLP signed-distance field with an eikonal regularizer; the surface is its zero level set (marching cubes).
13
 
14
+ Trained from scratch in **[Ropedia Academy](https://chaoyue0307.github.io/ropedia-academy/)** — an interactive, bilingual course on embodied & spatial AI. **Educational model:** small and quick to train; the value is the *method* and a reproducible pipeline, not a leaderboard score.
15
 
16
+ | | |
17
+ |---|---|
18
+ | **Task** | implicit 3D shape |
19
+ | **Data** | analytic torus samples |
20
+ | **Track** | B · 3D & rendering |
21
+ | **Notebook** | [![Open In Colab](https://colab.research.google.com/assets/colab-badge.svg)](https://colab.research.google.com/github/ChaoYue0307/ropedia-academy/blob/main/notebooks/training/B_deepsdf_shape.ipynb) |
22
+
23
+ ## Results
24
+
25
+ | metric | value |
26
+ |---|---|
27
+ | l1 (final) | 0.016 |
28
+
29
+
30
+ ![figure](figure.png)
31
+
32
+ ## How to use
33
+
34
+ ```python
35
+ import torch
36
+ state = torch.load("model.pt", map_location="cpu") # some labs save pose.pt / gaussians.pt / transform.pt
37
+ # Rebuild the model class from the Ropedia Academy notebook (linked above), then:
38
+ # model.load_state_dict(state)
39
+ ```
40
+
41
+ ## Files
42
+
43
+ - `figure.png`
44
+ - `metrics.json`
45
+ - `sdf.pt`
46
+
47
+
48
+ ## Reproduce / train your own
49
+
50
+ Open the [lab notebook in Colab](https://colab.research.google.com/github/ChaoYue0307/ropedia-academy/blob/main/notebooks/training/B_deepsdf_shape.ipynb) → **Runtime → GPU → Run all**, then its *Publish to the Hugging Face Hub* cell. Browse every lab in the [Ropedia Academy Labs tab](https://chaoyue0307.github.io/ropedia-academy/labs).
51
+
52
+
53
+ ---
54
+ *Part of the [Ropedia Academy](https://chaoyue0307.github.io/ropedia-academy/) trained-model collection.*
figure.png CHANGED

Git LFS Details

  • SHA256: 3fa66a72cf820515a7d18b70b943aeecb281ed02110eb1e02b7da2fe474fb0a5
  • Pointer size: 131 Bytes
  • Size of remote file: 126 kB

Git LFS Details

  • SHA256: f747d20e9742bcc737542512231ff01877ae4912d271db24c5c140a7df106235
  • Pointer size: 130 Bytes
  • Size of remote file: 69.5 kB
metrics.json CHANGED
@@ -2,47 +2,47 @@
2
  "l1": [
3
  [
4
  0,
5
- 0.09738403558731079
6
  ],
7
  [
8
  150,
9
- 0.004130750894546509
10
  ],
11
  [
12
  300,
13
- 0.003667763201519847
14
  ],
15
  [
16
  450,
17
- 0.00297424360178411
18
  ],
19
  [
20
  600,
21
- 0.00217777700163424
22
  ],
23
  [
24
  750,
25
- 0.002252743812277913
26
  ],
27
  [
28
  900,
29
- 0.0018030456267297268
30
  ],
31
  [
32
  1050,
33
- 0.0018078343709930778
34
  ],
35
  [
36
  1200,
37
- 0.0016836210852488875
38
  ],
39
  [
40
  1350,
41
- 0.0016222801059484482
42
  ],
43
  [
44
  1500,
45
- 0.0017965107690542936
46
  ]
47
  ]
48
  }
 
2
  "l1": [
3
  [
4
  0,
5
+ 0.0619
6
  ],
7
  [
8
  150,
9
+ 0.0245
10
  ],
11
  [
12
  300,
13
+ 0.0196
14
  ],
15
  [
16
  450,
17
+ 0.018
18
  ],
19
  [
20
  600,
21
+ 0.0191
22
  ],
23
  [
24
  750,
25
+ 0.0182
26
  ],
27
  [
28
  900,
29
+ 0.0176
30
  ],
31
  [
32
  1050,
33
+ 0.0168
34
  ],
35
  [
36
  1200,
37
+ 0.0167
38
  ],
39
  [
40
  1350,
41
+ 0.0157
42
  ],
43
  [
44
  1500,
45
+ 0.016
46
  ]
47
  ]
48
  }
sdf.pt CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:0586714031168a732b98f5c714e89fbc937407b283a971ba1749eca17885f958
3
  size 559189
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:5cd6f754d0b3e7805aff5e59b7e56c351b6d50f440ad6b08bee5a4b042728ced
3
  size 559189