Instructions to use torch-pointcloud/act.scanobjectnn-objonly.runpei-dong with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- torch-pointcloud
How to use torch-pointcloud/act.scanobjectnn-objonly.runpei-dong with torch-pointcloud:
import torch_pointcloud as tp model = tp.create_model("act.scanobjectnn-objonly.runpei-dong", pretrained=True) - Notebooks
- Google Colab
- Kaggle
Download config.json from torch-pointcloud/act.scanobjectnn-objonly.runpei-dong: direct link, hf CLI and curl.
- Browser
- Download file 927 Bytes
-
https://huggingface.co/torch-pointcloud/act.scanobjectnn-objonly.runpei-dong/resolve/main/config.json
- Command line
-
hf download hf://torch-pointcloud/act.scanobjectnn-objonly.runpei-dong/config.json
-
curl -L -o config.json https://huggingface.co/torch-pointcloud/act.scanobjectnn-objonly.runpei-dong/resolve/main/config.json
927 Bytes
| { | |
| "library_name": "torch-pointcloud", | |
| "name": "act.scanobjectnn-objonly.runpei-dong", | |
| "task": "classification", | |
| "hparams": { | |
| "in_channels": 0, | |
| "num_classes": 15, | |
| "embed_dim": 384, | |
| "depth": 12, | |
| "num_heads": 6, | |
| "num_groups": 128, | |
| "group_size": 32, | |
| "encoder_local_channels": [ | |
| 128, | |
| 256 | |
| ], | |
| "encoder_global_channels": [ | |
| 512 | |
| ], | |
| "pos_embed_channels": [ | |
| 128 | |
| ], | |
| "drop_path": 0.1, | |
| "dropout": 0.5, | |
| "act": "gelu", | |
| "spatial_dim": 3 | |
| }, | |
| "input_keys": [ | |
| "x", | |
| "pos", | |
| "batch" | |
| ], | |
| "dataset": "scanobjectnn-objonly", | |
| "metrics": { | |
| "OA": 91.22, | |
| "mAcc": 90.26 | |
| }, | |
| "classes": [ | |
| "bag", | |
| "bin", | |
| "box", | |
| "cabinet", | |
| "chair", | |
| "desk", | |
| "display", | |
| "door", | |
| "shelf", | |
| "table", | |
| "bed", | |
| "pillow", | |
| "sink", | |
| "sofa", | |
| "toilet" | |
| ], | |
| "author": "runpei-dong", | |
| "license": "MIT" | |
| } | |