!9731 Optimize RISE for better performance

From: @lixiaohui33
Reviewed-by: @wuxuejian,@liucunwei
Signed-off-by: @liucunwei

mindspore/explainer/_image_classification_runner.py

@@ -77,16 +77,16 @@ class ImageClassificationRunner:
         >>> from mindspore.train.serialization import load_checkpoint, load_param_into_net
         >>> # Prepare the dataset for explaining and evaluation, e.g., Cifar10
         >>> dataset = get_dataset('/path/to/Cifar10_dataset')
-        >>> labels = ['airplane', 'automobile', 'bird', 'cat', 'deer', 'dog', 'frog', 'horse', 'ship', 'turck']
+        >>> labels = ['airplane', 'automobile', 'bird', 'cat', 'deer', 'dog', 'frog', 'horse', 'ship', 'truck']
         >>> # load checkpoint to a network, e.g. checkpoint of resnet50 trained on Cifar10
         >>> param_dict = load_checkpoint("checkpoint.ckpt")
-        >>> net = resnet50(len(classes))
+        >>> net = resnet50(len(labels))
         >>> activation_fn = Softmax()
         >>> load_param_into_net(net, param_dict)
         >>> gbp = GuidedBackprop(net)
         >>> gradient = Gradient(net)
         >>> explainers = [gbp, gradient]
-        >>> faithfulness = Faithfulness(len(labels), "NaiveFaithfulness", activation_fn)
+        >>> faithfulness = Faithfulness(len(labels), activation_fn, "NaiveFaithfulness")
         >>> benchmarkers = [faithfulness]
         >>> runner = ImageClassificationRunner("./summary_dir", (dataset, labels), net, activation_fn)
         >>> runner.register_saliency(explainers=explainers, benchmarkers=benchmarkers)

mindspore/explainer/benchmark/_attribution/class_sensitivity.py

@@ -16,6 +16,7 @@
 
 import numpy as np
 
+from mindspore.explainer.explanation import RISE
 from .metric import LabelAgnosticMetric
 from ... import _operators as ops
 from ..._utils import calc_correlation
@@ -55,7 +56,7 @@ class ClassSensitivity(LabelAgnosticMetric):
             >>> # prepare your explainer to be evaluated, e.g., Gradient.
             >>> gradient = Gradient(network)
             >>> input_x = ms.Tensor(np.random.rand(1, 3, 224, 224), ms.float32)
-            >>> # class_sensitivity is a ClassSensitivity instance
+            >>> class_sensitivity = ClassSensitivity()
             >>> res = class_sensitivity.evaluate(gradient, input_x)
         """
         self._check_evaluate_param(explainer, inputs)
@@ -64,7 +65,12 @@ class ClassSensitivity(LabelAgnosticMetric):
 
         max_confidence_label = ops.argmax(outputs)
         min_confidence_label = ops.argmin(outputs)
-
+        if isinstance(explainer, RISE):
+            labels = ops.stack([max_confidence_label, min_confidence_label], axis=1)
+            full_saliency = explainer(inputs, labels)
+            max_confidence_saliency = full_saliency[:, max_confidence_label].asnumpy()
+            min_confidence_saliency = full_saliency[:, min_confidence_label].asnumpy()
+        else:
             max_confidence_saliency = explainer(inputs, max_confidence_label).asnumpy()
             min_confidence_saliency = explainer(inputs, min_confidence_label).asnumpy()
 

mindspore/explainer/explanation/_attribution/_perturbation/rise.py

@@ -14,11 +14,9 @@
 # ============================================================================
 """RISE."""
 import math
-import random
 
 import numpy as np
 
-from mindspore.ops.operations import Concat
 from mindspore import Tensor
 from mindspore.train._utils import check_value_type
 
@@ -107,18 +105,13 @@ class RISE(PerturbationAttribution):
         up_size = (height + mask_size[0], width + mask_size[1])
         mask = np.random.random((batch_size, 1) + mask_size) < self._mask_probability
         upsample = resize(op.Tensor(mask, data.dtype), up_size,
-                          self._resize_mode)
+                          self._resize_mode).asnumpy()
+        shift_x = np.random.randint(0, mask_size[0] + 1, size=batch_size)
+        shift_y = np.random.randint(0, mask_size[1] + 1, size=batch_size)
 
-        # Pack operator not available for GPU, thus transfer to numpy first
+        masks = [sample[:, x_i: x_i + height, y_i: y_i + width] for sample, x_i, y_i
-        masks_lst = []
+                 in zip(upsample, shift_x, shift_y)]
-        for sample in upsample:
+        masks = Tensor(np.array(masks), data.dtype)
-            shift_x = random.randint(0, mask_size[0])
-            shift_y = random.randint(0, mask_size[1])
-            masks_lst.append(sample[:, shift_x: shift_x + height, shift_y:shift_y + width])
-
-        concat = Concat()
-        masks = concat(tuple(masks_lst))
-        masks = op.reshape(masks, (batch_size, -1, height, width))
         return masks
 
     def __call__(self, inputs, targets):
@@ -157,11 +150,8 @@ class RISE(PerturbationAttribution):
 
         attr_np = attr_np / self._num_masks
         targets = self._unify_targets(inputs, targets)
-        attr_classes = []
+
-        for idx, target in enumerate(targets):
+        attr_classes = [att_i[target] for att_i, target in zip(attr_np, targets)]
-            attr_np_idx = attr_np[idx]
-            attr_idx = attr_np_idx[target]
-            attr_classes.append(attr_idx)
 
         return op.Tensor(attr_classes, dtype=inputs.dtype)