add example rdf.yaml

example_descriptions/models/unet2d_multi_tensor/rdf.yaml

@@ -0,0 +1,73 @@
+authors:
+  - { name: Constantin Pape, github_user: constantinpape }
+cite:
+  - { text: training library, doi: "10.5281/zenodo.5108853" }
+  - { text: architecture, doi: "10.1007/978-3-319-24574-4_28" }
+covers: [cover.jpg]
+dependencies: conda:environment.yaml
+description: Multi tensor
+documentation: documentation.md
+format_version: 0.4.0
+git_repo: https://github.com/constantinpape/torch-em.git
+inputs:
+  - axes: bcyx
+    data_range: [-.inf, .inf]
+    data_type: float32
+    name: input0
+    preprocessing:
+      - kwargs: { axes: cyx, mode: per_sample }
+        name: zero_mean_unit_variance
+    shape:
+      min: [1, 1, 32, 32]
+      step: [0, 0, 16, 16]
+  - axes: bcyx
+    data_range: [-.inf, .inf]
+    data_type: float32
+    name: input1
+    preprocessing:
+      - kwargs: { axes: cyx, mode: per_sample }
+        name: zero_mean_unit_variance
+    shape:
+      min: [1, 1, 32, 32]
+      step: [0, 0, 16, 16]
+license: CC-BY-4.0
+links: [ilastik/ilastik]
+name: Multi-tensor
+outputs:
+  - axes: bcyx
+    data_range: [-.inf, .inf]
+    data_type: float32
+    name: output0
+    shape:
+      offset: [0, 0, 0, 0]
+      reference_tensor: input0
+      scale: [1, 1, 1, 1]
+  - axes: bcyx
+    data_range: [-.inf, .inf]
+    data_type: float32
+    name: output1
+    shape:
+      offset: [0, 0, 0, 0]
+      reference_tensor: input1
+      scale: [1, 1, 1, 1]
+tags: [segmentation]
+test_inputs: [test_input_0.npy, test_input_1.npy]
+test_outputs: [test_output_0.npy, test_output_1.npy]
+timestamp: "2021-09-13T15:55:34.193995"
+type: model
+weights:
+  onnx:
+    opset_version: 12
+    sha256: 9b5bd88a3d29cf9979b30c03b4d5af12fdfa1d7193f5d2f2cc3942ffcf71ce3c
+    source: weights.onnx
+    parent: pytorch_state_dict
+  torchscript:
+    sha256: 097bb5062df1fe48a5e7473ea2f6025c77d334a9e3f92af79fc3d6d530c01720
+    source: weights-torchscript.pt
+    parent: pytorch_state_dict
+  pytorch_state_dict:
+    architecture: multi_tensor_unet.py:MultiTensorUNet
+    architecture_sha256: 5e3d36b5187b85d5c935f2efde7cafe293dbffa413618f49a0744bf1be75c22b
+    kwargs: { depth: 3, in_channels: 2, initial_features: 16, out_channels: 2 }
+    sha256: c498522b3f2b02429b41fe9dbcb722ce0d7ad4cae7fcf8059cee27857ae49b00
+    source: weights.pt