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
Copy file name to clipboardExpand all lines: dev/_sources/overview/EDITING.md.txt
+1-1Lines changed: 1 addition & 1 deletion
Original file line number
Diff line number
Diff line change
@@ -5,7 +5,7 @@ However, for your convenience, we allow manual edits in various stages of the Fa
5
5
These editing options include approaches that are inherited from FreeSurfer as well as some FastSurfer-specific editing options.
6
6
7
7
The provided editing options may be changed or extended in the future, also depending on requests from the community.
8
-
Furthermore, we invite users to [contribute](CONTRIBUTING.md) such changes and/or datasets of paired MRI images and edited files to improve FastSurfer's neural networks.
8
+
Furthermore, we invite users to [contribute](../../CONTRIBUTING.md) such changes and/or datasets of paired MRI images and edited files to improve FastSurfer's neural networks.
Get FastSurfer from GitHub. Here you can decide if you want to install the current experimental "dev" version (which can be broken) or the "stable" branch (that has been tested thoroughly):
86
85
87
86
```bash
87
+
cd /path/to/install
88
+
# FastSurfer will get cloned to /path/to/install/FastSurfer
Create a new environment and install FastSurfer dependencies:
95
96
96
97
```bash
97
-
conda env create -f ./env/fastsurfer.yml
98
-
conda activate fastsurfer
98
+
# make sure you are in the FastSurfer directory!
99
+
# create a .venv environment directory inside /path/to/install/FastSurfer with the FastSurfer dependencies
100
+
# the minimum required python version is 3.10
101
+
uv venv --python python3.12
102
+
# download and install packages for the fastsurfer environment (implicitly read from requirements.txt)
103
+
uv pip sync pyproject.toml
104
+
```
105
+
`uv` will also try to find the correct backend for your hardware, but you can manually specify the backend for testing:
106
+
purposes:
107
+
```bash
108
+
# make sure you are in the FastSurfer directory!
109
+
uv pip sync pyproject.toml --torch-backend cpu
110
+
```
111
+
You can now activate the FastSurfer environment with
112
+
```bash
113
+
source .venv/bin/activate
99
114
```
100
115
101
-
If you do not have an NVIDIA GPU, you can create appropriate ymls on the fly with `python ./tools/Docker/install_env.py -m $MODE -i ./env/FastSurfer.yml -o ./fastsurfer_$MODE.yml`. Here `$MODE` can be for example `cpu`, see also `python ./tools/Docker/install_env.py --help` for other options like rocm or cuda versions. Finally, replace `./env/fastsurfer.yml` with your custom environment file `./fastsurfer_$MODE.yml`.
102
-
If you only want to run the surface pipeline, use `./env/fastsurfer_reconsurf.yml`.
103
-
104
-
Next, add the fastsurfer directory to the python path (make sure you have changed into it already):
116
+
Next, add the fastsurfer directory to the python path:
105
117
```bash
118
+
# make sure you are in the FastSurfer directory!
106
119
export PYTHONPATH="${PYTHONPATH}:$PWD"
107
120
```
108
121
109
122
This will need to be done every time you want to run FastSurfer, or you need to add this line to your `~/.bashrc` if you are using bash, for example:
Once all dependencies are installed, you are ready to run the FastSurfer segmentation-only (!!) pipeline by calling ```./run_fastsurfer.sh --seg_only ....``` , see [Example 3](EXAMPLES.md#example-3-native-fastsurfer-on-subjectx-with-parallel-processing-of-hemis) for command line flags.
@@ -177,15 +192,15 @@ To install brew and then python3.10, execute the following in a Terminal:
0 commit comments