Safran/KPConv-PyTorch
1
0
Fork 0
Code Issues Pull requests Packages Projects Releases Wiki Activity

Adding pretrained weights S3DIS

Browse source
This commit is contained in:
HuguesTHOMAS 2021-08-02 13:31:31 +00:00
parent 3d683b6bd6
commit d1bb1ca36e
4 changed files with 51 additions and 8 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

47
doc/pretrained_models_guide.md
View file

@ -1,5 +1,48 @@
## Test a pretrained network
## S3DIS Pretrained Models
### Models
We provide pretrained weights for S3DIS dataset. The raw weights come with a parameter file describing the architecture and network hyperparameters. THe code can thus load the network automatically.
The instructions to run these models are in the S3DIS documentation, section [Test the trained model](./doc/scene_segmentation_guide.md#test-the-trained-model).
| Name (link) | KPConv Type | Description | Score |
|:-------------|:-------------:|:-----|:-----:|
| [Light_KPFCNN](https://drive.google.com/file/d/14sz0hdObzsf_exxInXdOIbnUTe0foOOz/view?usp=sharing) | rigid | A network with small `in_radius` for light GPU consumption (~8GB) | 65.4% |
| [Heavy_KPFCNN](https://drive.google.com/file/d/1ySQq3SRBgk2Vt5Bvj-0N7jDPi0QTPZiZ/view?usp=sharing) | rigid | A network with better performances but needing bigger GPU (>18GB). | 66.4% |
### Instructions
1. Unzip and place the folder in your 'results' folder.
2. In the test script `test_any_model.py`, set the variable `chosen_log` to the path were you placed the folder.
3. Run the test script
python3 test_any_model.py
4. You will see the performance (on the subsampled input clouds) increase as the test goes on.
Confusion on sub clouds
65.08 | 92.11 98.40 81.83 0.00 18.71 55.41 68.65 90.93 79.79 74.83 65.31 63.41 56.62
5. After a few minutes, the script will reproject the results form the subsampled input clouds to the real data and get you the real score
Reproject Vote #9
Done in 2.6 s
Confusion on full clouds
Done in 2.1 s
--------------------------------------------------------------------------------------
65.38 | 92.62 98.39 81.77 0.00 18.87 57.80 67.93 91.52 80.27 74.24 66.14 64.01 56.42
--------------------------------------------------------------------------------------
6. The test script creates a folder `test/name-of-your-log`, where it saves the predictions, potentials, and probabilities per class. You can load them with CloudCompare for visualization.
TODO

2
plot_convergence.py
View file

@ -221,7 +221,7 @@ def compare_trainings(list_of_paths, list_of_labels=None):
print(path)
if ('val_IoUs.txt' in [f.decode('ascii') for f in listdir(path)]) or ('val_confs.txt' in [f.decode('ascii') for f in listdir(path)]):
if ('val_IoUs.txt' in [f for f in listdir(path)]) or ('val_confs.txt' in [f for f in listdir(path)]):
config = Config()
config.load(path)
else:

4
test_models.py
View file

@ -95,10 +95,10 @@ if __name__ == '__main__':
# > 'last_XXX': Automatically retrieve the last trained model on dataset XXX
# > '(old_)results/Log_YYYY-MM-DD_HH-MM-SS': Directly provide the path of a trained model
chosen_log = 'results/Log_2020-04-05_19-19-20' # => ModelNet40
chosen_log = 'results/Light_KPFCNN'
# Choose the index of the checkpoint to load OR None if you want to load the current checkpoint
chkp_idx = None
chkp_idx = -1
# Choose to test on validation or test split
on_val = True

6
utils/tester.py
View file

@ -360,10 +360,10 @@ class ModelTester:
proj_probs = []
for i, file_path in enumerate(test_loader.dataset.files):
print(i, file_path, test_loader.dataset.test_proj[i].shape, self.test_probs[i].shape)
# print(i, file_path, test_loader.dataset.test_proj[i].shape, self.test_probs[i].shape)
print(test_loader.dataset.test_proj[i].dtype, np.max(test_loader.dataset.test_proj[i]))
print(test_loader.dataset.test_proj[i][:5])
# print(test_loader.dataset.test_proj[i].dtype, np.max(test_loader.dataset.test_proj[i]))
# print(test_loader.dataset.test_proj[i][:5])
# Reproject probs on the evaluations points
probs = self.test_probs[i][test_loader.dataset.test_proj[i], :]