Instructions to use torch-pointcloud/pointnext-base.s3dis-area4.openpoints with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- torch-pointcloud
How to use torch-pointcloud/pointnext-base.s3dis-area4.openpoints with torch-pointcloud:
import torch_pointcloud as tp model = tp.create_model("pointnext-base.s3dis-area4.openpoints", pretrained=True) - Notebooks
- Google Colab
- Kaggle
Update config.json for pointnext-base.s3dis-area4.openpoints
Browse files- config.json +95 -0
config.json
ADDED
|
@@ -0,0 +1,95 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"library_name": "torch-pointcloud",
|
| 3 |
+
"name": "pointnext-base.s3dis-area4.openpoints",
|
| 4 |
+
"task": "semantic-segmentation",
|
| 5 |
+
"hparams": {
|
| 6 |
+
"in_channels": 4,
|
| 7 |
+
"num_classes": 13,
|
| 8 |
+
"spatial_dim": 3,
|
| 9 |
+
"stem_plain_last": true,
|
| 10 |
+
"decoder_plain_last": false,
|
| 11 |
+
"ratios": [
|
| 12 |
+
0.25,
|
| 13 |
+
0.25,
|
| 14 |
+
0.25,
|
| 15 |
+
0.25
|
| 16 |
+
],
|
| 17 |
+
"radiuses": [
|
| 18 |
+
0.1,
|
| 19 |
+
0.2,
|
| 20 |
+
0.4,
|
| 21 |
+
0.8,
|
| 22 |
+
1.6
|
| 23 |
+
],
|
| 24 |
+
"num_neighbors": [
|
| 25 |
+
32,
|
| 26 |
+
32,
|
| 27 |
+
32,
|
| 28 |
+
32,
|
| 29 |
+
32
|
| 30 |
+
],
|
| 31 |
+
"act": "relu",
|
| 32 |
+
"act_first": false,
|
| 33 |
+
"norm": "batch_norm",
|
| 34 |
+
"bias": true,
|
| 35 |
+
"add_self_loops": false,
|
| 36 |
+
"stem_channels": 32,
|
| 37 |
+
"encoder_channels": [
|
| 38 |
+
64,
|
| 39 |
+
128,
|
| 40 |
+
256,
|
| 41 |
+
512
|
| 42 |
+
],
|
| 43 |
+
"encoder_depths": [
|
| 44 |
+
1,
|
| 45 |
+
2,
|
| 46 |
+
1,
|
| 47 |
+
1
|
| 48 |
+
],
|
| 49 |
+
"encoder_expansion": 4,
|
| 50 |
+
"sa_layers": 1,
|
| 51 |
+
"sa_use_res": false,
|
| 52 |
+
"decoder_channels": [
|
| 53 |
+
256,
|
| 54 |
+
128,
|
| 55 |
+
64,
|
| 56 |
+
32
|
| 57 |
+
],
|
| 58 |
+
"decoder_depths": [
|
| 59 |
+
2,
|
| 60 |
+
2,
|
| 61 |
+
2,
|
| 62 |
+
2
|
| 63 |
+
],
|
| 64 |
+
"head_channels": [
|
| 65 |
+
32
|
| 66 |
+
]
|
| 67 |
+
},
|
| 68 |
+
"input_keys": [
|
| 69 |
+
"x",
|
| 70 |
+
"pos",
|
| 71 |
+
"batch"
|
| 72 |
+
],
|
| 73 |
+
"dataset": "s3dis-area4",
|
| 74 |
+
"metrics": {
|
| 75 |
+
"mIoU": 62.63,
|
| 76 |
+
"OA": 87.03
|
| 77 |
+
},
|
| 78 |
+
"classes": [
|
| 79 |
+
"ceiling",
|
| 80 |
+
"floor",
|
| 81 |
+
"wall",
|
| 82 |
+
"beam",
|
| 83 |
+
"column",
|
| 84 |
+
"window",
|
| 85 |
+
"door",
|
| 86 |
+
"chair",
|
| 87 |
+
"table",
|
| 88 |
+
"bookcase",
|
| 89 |
+
"sofa",
|
| 90 |
+
"board",
|
| 91 |
+
"clutter"
|
| 92 |
+
],
|
| 93 |
+
"author": "openpoints",
|
| 94 |
+
"license": "MIT"
|
| 95 |
+
}
|