added model card and model weights

README.md

@@ -0,0 +1,84 @@
+# NCPDNet-singlecoil-radial
+---
+tags:
+- TensorFlow
+- MRI reconstruction
+- MRI
+datasets:
+- fastMRI
+---
+
+This is a non-Cartesian MRI reconstruction model for radial trajectories at acceleration factor 4.
+The model uses 10 iterations and a small vanilla CNN.
+
+## Model description
+For more details, see https://hal.inria.fr/hal-03188997.
+This section is WIP.
+
+## Intended uses and limitations
+This model can be used to reconstruct knee data from Siemens scanner at acceleration factor 4 in a radial acquisition setting.
+
+## How to use
+This model can be loaded using the following repo: https://github.com/zaccharieramzi/fastmri-reproducible-benchmark.
+After cloning the repo, `git clone https://github.com/zaccharieramzi/fastmri-reproducible-benchmark`, you can install the package via `pip install fastmri-reproducible-benchmark`.
+The framework is TensorFlow.
+
+You can initialize and load the model weights as follows:
+```python
+import tensorflow as tf
+
+from fastmri_recon.models.subclassed_models.ncpdnet import NCPDNet
+
+
+model = NCPDNet(
+    im_size=(640, 400),
+    dcomp=True,
+)
+kspace_shape = 1
+inputs = [
+    tf.zeros([1, 1, kspace_shape, 1], dtype=tf.complex64),
+    tf.zeros([1, 2, kspace_shape], dtype=tf.float32),
+    (tf.constant([320]), tf.ones([1, kspace_shape], dtype=tf.float32)),
+]
+model(inputs)
+model.load_weights('model_weights.h5')
+```
+
+Using the model is then as simple as:
+```python
+model([
+    kspace,  # shape: [n_slices, 1, n_kspace_samples, 1]
+    traj,  # shape: [n_slices, 1, 2, n_kspace_samples]
+    (
+        output_shape,  # shape: [n_slices, 1]
+        dcomp,  # shape: [n_slices, n_kspace_samples]
+    )
+])
+```
+
+## Limitations and bias
+The limitations and bias of this model have not been properly investigated.
+
+## Training data
+This model was trained using the [fastMRI dataset](https://fastmri.org/dataset/).
+
+## Training procedure
+The training procedure is described in https://hal.inria.fr/hal-03188997.
+This section is WIP.
+
+## Evaluation results
+On the fastMRI validation dataset:
+- PSNR: 32.66
+- SSIM: 0.7327
+
+
+## Bibtex entry
+```
+@unpublished{ramzi:hal-03188997,
+  TITLE = {{NC-PDNet: a Density-Compensated Unrolled Network for 2D and 3D non-Cartesian MRI Reconstruction}},
+  AUTHOR = {Ramzi, Zaccharie and G R, Chaithya and Starck, Jean-Luc and Ciuciu, Philippe},
+  YEAR = {2021},
+  MONTH = Sep,
+}
+```
+

model_weights.h5

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:d6f36293e0535c9722781cacfe71dcceda925b5ea23763a398921fb8e504e55d
+size 716248