Merge pull request #16 from brudfors/fix-dependencies

Fix dependencies

README.md

@@ -23,18 +23,27 @@ An installation-free demo of UniRes is available in Colab:
 
 ## 1. Python
 
-### 1.1. Dependencies
-The *NITorch* package is required to use *UniRes*; simply follow the quickstart 
-guide on its GitHub page: https://github.com/balbasty/nitorch.
-
-Next, activate a *NITorch* virtual environment and move to the *UniRes* project 
-directory. Then install the package using either setuptools or pip:
-``` bash
-cd /path/to/unires
+### 1.1. Installation
+Clone `UniRes`:
+```shell
+git clone https://github.com/brudfors/UniRes
+```
+Then `cd` into the `UniRes` folder and install it by:
+```shell
 pip install .
-``` 
+```
+**OBS**: The algorithm runs much faster if the compiled backend is used:
+```shell
+NI_COMPILED_BACKEND="C" pip install --no-build-isolation .
+```
+However, for running on the GPU, this only works if you ensure that the PyTorch installation uses the same CUDA version that is on your system; therefore, it might be worth installing PyTorch beforehand, *i.e.*:
+```shell
+pip install torch==1.9.0+cu111
+NI_COMPILED_BACKEND="C" pip install --no-build-isolation .
+```
+where the PyTorch CUDA version matches the output of `nvcc --version`.
 
-### 1.2. Example use cases
+### 1.2. Example usage
 
 Running *UniRes* is straight forward. Let's say you have three 
 MR images: `T1.nii.gz`, `T2.nii.gz` and `PD.nii.gz`, then 

setup.py

@@ -1,15 +1,17 @@
-from setuptools import setup, find_packages
+from setuptools import (setup, find_packages)
 
 
 setup(
-    name='unires',
-    version='0.1a',
-    packages=find_packages(),
-    url='https://github.com/brudfors/UniRes',
     author='Mikael Brudfors',
     author_email='[email protected]',
+    description='UniRes: Unified Super-Resolution of Medical Imaging Data',    
     entry_points={'console_scripts': ['unires=unires._cli:run']},
-    description='UniRes: Unified Super-Resolution of Neuroimaging Data',
-    python_requires='>=3.6'
-
+    install_requires=[
+        "nitorch[all]@git+https://github.com/balbasty/[email protected]#egg=nitorch",
+    ],
+    name='unires',    
+    packages=find_packages(),
+    python_requires='>=3.6',
+    url='https://github.com/brudfors/UniRes',
+    version='0.0.1a',        
 )

unires/run.py

@@ -238,7 +238,8 @@ def init(data, sett=settings()):
             # Makes recons aligned with same grid, across subjects
             sett.do_atlas_align = True
             sett.crop = True
-            sett.pow = 256
+            if sett.pow == 0:
+                sett.pow = 256
 
         # Read and format data (images and labels)
         x = _read_data(data, sett)