Instructions to use torch-pointcloud/concerto-tiny.pretrain.pointcept with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- torch-pointcloud
How to use torch-pointcloud/concerto-tiny.pretrain.pointcept with torch-pointcloud:
import torch_pointcloud as tp model = tp.create_model("concerto-tiny.pretrain.pointcept", pretrained=True) - Notebooks
- Google Colab
- Kaggle
File size: 1,163 Bytes
e1bc417 | 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 | {
"library_name": "torch-pointcloud",
"name": "concerto-tiny.pretrain.pointcept",
"task": "pretraining",
"hparams": {
"in_channels": 9,
"serialization_orders": [
"z",
"z-trans",
"hilbert",
"hilbert-trans"
],
"shuffle_serialization_orders": true,
"strides": [
2,
2,
2,
2
],
"encoder_depths": [
1,
1,
1,
3,
1
],
"encoder_channels": [
16,
32,
64,
128,
256
],
"encoder_num_heads": [
1,
2,
4,
8,
16
],
"encoder_patch_size": [
1024,
1024,
1024,
1024,
1024
],
"norm": "layer_norm",
"act": "gelu",
"mlp_ratio": 4.0,
"qkv_bias": true,
"qk_scale": null,
"attn_drop": 0.0,
"proj_drop": 0.0,
"drop_path": 0.3,
"use_flash_attn": true,
"upcast_attn": false,
"upcast_softmax": false,
"pooling": "grid",
"stem_type": "linear",
"norm_kwargs": {
"mode": "node"
}
},
"input_keys": [
"x",
"pos_grid",
"batch"
],
"author": "pointcept",
"license": "CC-BY-NC-4.0"
}
|