A_dagger integration in benchmark for NBUdataset

deepinv_bench/benchmarks/nbu_pansharpening/solvers/brovey.py

@@ -0,0 +1,27 @@
+from benchopt import BaseSolver
+
+import torch
+import deepinv as dinv
+
+
+class Brovey(dinv.models.Reconstructor):
+    def forward(self, y, physics):
+        return physics.A_dagger(y)
+
+
+class Solver(BaseSolver):
+    name = "Brovey"
+
+    parameters = {}
+
+    def set_objective(self, train_dataset=None, physics=None):
+        device = dinv.utils.get_freer_gpu() if torch.cuda.is_available() else "cpu"
+
+        self.model = Brovey().to(device)
+        self.model.device = device
+
+    def run(self, _):
+        pass
+
+    def get_result(self):
+        return dict(model=self.model)