Refactor module eval to evaluate

Author: spencerwooo

README.md

@@ -83,7 +83,7 @@ attack_args = {'steps': 10, 'decay': 1.0}
 attack = create_attack('MIFGSM', model, normalize, device, eps=0.03, **attack_args)
 ```
 
-Check out [examples/](examples/mifgsm_transfer.py) and [`torchattack.eval.runner`](torchattack/eval/runner.py) for full examples.
+Check out [examples/](examples/mifgsm_transfer.py) and [`torchattack.evaluate.runner`](torchattack/evaluate/runner.py) for full examples.
 
 ## Attacks
 

docs/api/eval/eval-dataset.md

@@ -0,0 +1,3 @@
+# NIPS 2017 Dataset
+
+::: torchattack.evaluate.dataset

docs/api/eval/eval-metric.md

@@ -0,0 +1,3 @@
+# Fooling rate metric
+
+::: torchattack.evaluate.metric

docs/api/eval/eval-runner.md

@@ -0,0 +1,3 @@
+# Attack runner
+
+::: torchattack.evaluate.runner

docs/api/eval/index.md

@@ -16,4 +16,4 @@ As a rule of thumb, we recommend saving images as PNGs, as they better keep the
 
 A commonly used perturbation magnitude is $\varepsilon = 8 / 255$, for reference.
 
-::: torchattack.eval.save_image_batch
+::: torchattack.evaluate.save_image_batch

docs/api/evaluate/eval-dataset.md

@@ -0,0 +1,11 @@
+# Evaluation module
+
+!!! warning
+    This module has been renamed to `torchattacks.evaluate` as of [v1.5.1](https://github.com/spencerwooo/torchattack/releases/tag/v1.5.1). The previous `torchattack.eval` will be deprecated in the future.
+
+The evaluation module `torchattack.evaluate` provides a few tools that are commonly used in transfer-based attack experiments.
+
+- [Saving adversarial examples](eval-save-image.md)
+- [Fooling rate metric](eval-metric.md)
+- [NIPS 2017 Dataset](eval-dataset.md)
+- [Attack runner](eval-runner.md)