You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Benchmarks of `MLSTRUCT-FP<https://github.com/MLSTRUCT/MLSTRUCT-FP>`_ dataset.
18
+
Benchmarks of `MLStructFP<https://github.com/MLSTRUCT/MLStructFP>`_ dataset.
19
19
20
20
21
21
Description
22
22
-----------
23
23
24
24
This repo contains the segmentation and vectorization models for processing our
25
-
`MLSTRUCT-FP dataset <https://github.com/MLSTRUCT/MLSTRUCT-FP>`_. See the following
25
+
`MLStructFP dataset <https://github.com/MLSTRUCT/MLStructFP>`_. See the following
26
26
jupyter notebook files for more information and a quick start (in order):
27
27
28
-
- `create_data <https://github.com/MLSTRUCT/MLSTRUCT-FP_benchmarks/blob/master/create_data.ipynb>`_: Creates a dataset, assembles crops, and export data session
29
-
- `fp_unet <https://github.com/MLSTRUCT/MLSTRUCT-FP_benchmarks/blob/master/fp_unet.ipynb>`_: Creates U-Net model for wall segmentation
30
-
- `vectorization <https://github.com/MLSTRUCT/MLSTRUCT-FP_benchmarks/blob/master/vectorization.ipynb>`_: Vectorizes a model using Egiazarian et al. method
28
+
- `create_data <https://github.com/MLSTRUCT/MLStructFP_benchmarks/blob/master/create_data.ipynb>`_: Creates a dataset, assembles crops, and export data session
29
+
- `fp_unet <https://github.com/MLSTRUCT/MLStructFP_benchmarks/blob/master/fp_unet.ipynb>`_: Creates U-Net model for wall segmentation
30
+
- `vectorization <https://github.com/MLSTRUCT/MLStructFP_benchmarks/blob/master/vectorization.ipynb>`_: Vectorizes a model using Egiazarian et al. method
31
31
32
32
The weights for the best model (no_rot_256_50) can be downloaded at
33
33
`this link <https://drive.google.com/file/d/15ufkjoWOFyT0Cm-MEc9zQJCDJIooOgh7/view?usp=sharing>`_. For the vectorization model, follow the following links
@@ -44,8 +44,8 @@ To use this code, you need a Python 3.8 installation. Then, run the following se
0 commit comments