pointcept-bot commited on
Commit
014a6b8
·
verified ·
1 Parent(s): 9e16b42

Name the weights after their experiments (Pointcept config naming)

Browse files
README.md CHANGED
@@ -22,11 +22,12 @@ at a photo's native resolution, up to 100 MP.
22
 
23
  | File | Model | Training data | Size | sha256 |
24
  |---|---|---|---|---|
25
- | `l-dpt-l-zs.pth` | L-DPT-L zero-shot | 8 synthetic / multi-view sets (below) | 1.08 GB | `b4e7e95301fda7557760f787e7992f4ccca3c03cb70e46eadeb48e227b1b31ca` |
26
- | `l-dpt-l-mix.pth` | L-DPT-L mixed | the same + NYUv2, KITTI, DIODE train | 1.08 GB | `4852b823c68fea30b3360a249e71daf72a9c6c2c74f747602a117b42f35d36fa` |
27
 
28
- Each file holds the decoder only (fp32 weights, 271M parameters) and its model
29
- configuration. **The DINOv3 encoder is not included**: it is Meta's, under the DINOv3
 
30
  License. Get the ViT-7B/16 (LVD-1689M) weights from Meta: the original file
31
  `dinov3_vit7b16_pretrain_lvd1689m-a955f4ea.pth` of the DINOv3 release, or the gated
32
  Hugging Face repo
@@ -49,7 +50,7 @@ With the `ldpt` package of the L-DPT code release:
49
  from huggingface_hub import hf_hub_download
50
  from ldpt.models.l_dpt import api
51
 
52
- decoder = hf_hub_download("pointcept-bot/L-DPT", "l-dpt-l-zs.pth")
53
  dinov3 = "dinov3_vit7b16_pretrain_lvd1689m-a955f4ea.pth" # Meta's file (or the HF folder)
54
  model = api.load(decoder, encoder_weights=dinov3) # cuda if available
55
  depth = api.predict(model, image) # HxWx3 uint8 RGB -> HxW float32
@@ -80,8 +81,8 @@ DIODE is the combined indoor + outdoor validation set.
80
 
81
  | Model | NYUv2 | KITTI | ETH3D | DIODE |
82
  |---|---|---|---|---|
83
- | `l-dpt-l-zs.pth` | 3.55 / 98.30 | 5.98 / 97.85 | 3.75 / 99.01 | 19.34 / 85.65 |
84
- | `l-dpt-l-mix.pth` | 2.89 / 99.08 | 5.39 / 98.38 | 3.57 / 98.94 | 14.98 / 88.17 |
85
 
86
  NYUv2, KITTI and DIODE are zero-shot for the zero-shot model and in-domain for the mixed
87
  one. On photos in the wild, the zero-shot model separates thin foreground (foliage,
 
22
 
23
  | File | Model | Training data | Size | sha256 |
24
  |---|---|---|---|---|
25
+ | `depth-l-dpt-v1m1-0-l-zs.pth` | L-DPT-L zero-shot | 8 synthetic / multi-view sets (below) | 1.08 GB | `b4e7e95301fda7557760f787e7992f4ccca3c03cb70e46eadeb48e227b1b31ca` |
26
+ | `depth-l-dpt-v1m1-1-l-mix.pth` | L-DPT-L mixed | the same + NYUv2, KITTI, DIODE train | 1.08 GB | `4852b823c68fea30b3360a249e71daf72a9c6c2c74f747602a117b42f35d36fa` |
27
 
28
+ Each file is named after the experiment (config) that produced it
29
+ (`configs/hybrid/<name>.py` of the code release) and holds the decoder only (fp32
30
+ weights, 271M parameters) and its model configuration. **The DINOv3 encoder is not included**: it is Meta's, under the DINOv3
31
  License. Get the ViT-7B/16 (LVD-1689M) weights from Meta: the original file
32
  `dinov3_vit7b16_pretrain_lvd1689m-a955f4ea.pth` of the DINOv3 release, or the gated
33
  Hugging Face repo
 
50
  from huggingface_hub import hf_hub_download
51
  from ldpt.models.l_dpt import api
52
 
53
+ decoder = hf_hub_download("pointcept-bot/L-DPT", "depth-l-dpt-v1m1-0-l-zs.pth")
54
  dinov3 = "dinov3_vit7b16_pretrain_lvd1689m-a955f4ea.pth" # Meta's file (or the HF folder)
55
  model = api.load(decoder, encoder_weights=dinov3) # cuda if available
56
  depth = api.predict(model, image) # HxWx3 uint8 RGB -> HxW float32
 
81
 
82
  | Model | NYUv2 | KITTI | ETH3D | DIODE |
83
  |---|---|---|---|---|
84
+ | `depth-l-dpt-v1m1-0-l-zs.pth` | 3.55 / 98.30 | 5.98 / 97.85 | 3.75 / 99.01 | 19.34 / 85.65 |
85
+ | `depth-l-dpt-v1m1-1-l-mix.pth` | 2.89 / 99.08 | 5.39 / 98.38 | 3.57 / 98.94 | 14.98 / 88.17 |
86
 
87
  NYUv2, KITTI and DIODE are zero-shot for the zero-shot model and in-domain for the mixed
88
  one. On photos in the wild, the zero-shot model separates thin foreground (foliage,
l-dpt-l-zs.pth → depth-l-dpt-v1m1-0-l-zs.pth RENAMED
File without changes
l-dpt-l-mix.pth → depth-l-dpt-v1m1-1-l-mix.pth RENAMED
File without changes