init

ml-stat-Sustech · Dec 6, 2023 · eff90be · eff90be
1 parent e0edf60
commit eff90be
Show file tree

Hide file tree

Showing 78 changed files with 21 additions and 10,592 deletions.
diff --git a/.gitignore b/.gitignore
@@ -9,8 +9,8 @@
 __pycache__/
 *.pdf
 *.npz
-advertorch/__pycache__/
-advertorch.egg-info
+deepcp/__pycache__/
+deepcp.egg-info
 data/
 .pytest_cache/
 .cache/

diff --git a/CONTRIBUTING.md b/CONTRIBUTING.md
@@ -1,6 +1,6 @@
-# Contributing to AdverTorch
+# Contributing to deepcp
 
-Thank you considering contributing to AdverTorch!
+Thank you considering contributing to deepcp!
 
 This document provide brief guidelines for potential contributors.
 
@@ -10,19 +10,7 @@ We ask that you follow the `PEP8` coding style in your pull requests, [`flake8`]
 
 ---
 ### Detailed guidelines for contributing new attacks
-- *(mandatory)* The implementation file should be added to the folder `advertorch/attacks`, and the class should be imported in `advertorch/attacks/__init__.py`.
-- *(mandatory)* The attack should be included in different unit tests, this can be done by adding the attack class to different lists in `advertorch/test_utils.py`
-    + add to `general_input_attacks` if it can perturb input tensor of any shape (not limited to images),
-    + add to `image_only_attacks` if it only works on images,
-    + add to `label_attacks` if the attack manipulates labels,
-    + add to `feature_attacks` if the attack manipulates features,
-    + add to `batch_consistent_attacks` if the attack's behavior should be the same when it is applied to a single example or a batch,
-    + add to `targeted_only_attacks` if the attack is a label attack and does not work for the untargeted case,
-    + add entry to `attack_kwargs` in `advertorch/tests/test_attacks_running.py`, for setting the hyperparameters used for test.
-- *(mandatory)* Benchmark the attack with at least one performance measure, by adding a script to `advertorch_examples/attack_benchmarks`.
-- *(mandatory)* If the contributor has a GPU computer, run `pytest` locally to make sure all the tests pass. (This is because travis-ci currently do not provide GPU machines for continuous integration.) If the contributor does not have a GPU computer, please let us know in the pull request. 
-- *(optional)* When an attack can be compared against other implementations, a comparison test could be added to `advertorch/external_tests`.
-- *(optional)* Add an ipython notebook example.
+
 
 ---
 ### Copyright notice at the beginning of files

diff --git a/MANIFEST.in b/MANIFEST.in
@@ -1,8 +1,7 @@
 include tests/*.py
 include external_tests/*.py
-include advertorch/VERSION
+include deepcp/VERSION
 include pytest.ini
 include LICENSE
 include LICENSE.GPL
-include advertorch_examples/trained_models/*.pt
-include advertorch_examples/*.ipynb
+include deepcp_examples/*.ipynb
diff --git a/README.md b/README.md
@@ -1,23 +1,13 @@
-<!-- <a href="https://github.com/borealisai/advertorch" target="_blank"><img src="https://raw.githubusercontent.com/borealisai/advertorch/master/assets/logo.png?raw=true" alt="advertorch logo" width="200"></a>
-
-[![Build Status](https://travis-ci.org/BorealisAI/advertorch.svg?branch=master)](https://travis-ci.org/BorealisAI/advertorch)
-
-
-<a href="https://github.com/borealisai/advertorch" target="_blank"><img src="https://raw.githubusercontent.com/borealisai/advertorch/master/assets/advertorch.png?raw=true" alt="advertorch text" width="100"></a>  -->
-
-DeepCP is a Python toolbox for conformal prediction research. The primary functionalities are implemented in PyTorch. Specifically, DeepCP contains modules of post-hoc methods and training methods for classification problems and regression problems.
-
-
-#### Latest version (v0.1)
+DeepCP is a Python toolbox for conformal prediction research on deep learning models. The primary functionalities are implemented in PyTorch. Specifically, DeepCP contains modules of post-hoc methods and training methods for classification problems and regression problems.
 
 ## Installation
 
-### Installing AdverTorch itself
+### Installing DeepCP itself
 
-We developed DeepCP under Python 3.8 and PyTorch 1.0.0 & 0.4.1. To install DeepCP, simply run
+We developed DeepCP under Python 3.9 and PyTorch 2.0.1. To install DeepCP, simply run
 
 ```
-pip install deeptorch
+pip install deepcp
 ```
 
 or clone the repo and run
@@ -30,39 +20,9 @@ To install the package in "editable" mode:
 pip install -e .
 ```
 
-### Setting up the testing environments
-
-Some attacks are tested against implementations in [Foolbox](https://github.com/bethgelab/foolbox) or [CleverHans](https://github.com/tensorflow/cleverhans) to ensure correctness. Currently, they are tested under the following versions of related libraries.
-```
-conda install -c anaconda tensorflow-gpu==1.11.0
-pip install git+https://github.com/tensorflow/cleverhans.git@336b9f4ed95dccc7f0d12d338c2038c53786ab70
-pip install Keras==2.2.2
-pip install foolbox==1.3.2
-```
-
 
 ## Examples
 ```python
-# prepare your pytorch model as "model"
-# prepare a batch of data and label as "cln_data" and "true_label"
-# ...
-
-from advertorch.attacks import LinfPGDAttack
-
-adversary = LinfPGDAttack(
-    model, loss_fn=nn.CrossEntropyLoss(reduction="sum"), eps=0.3,
-    nb_iter=40, eps_iter=0.01, rand_init=True, clip_min=0.0, clip_max=1.0,
-    targeted=False)
-
-adv_untargeted = adversary.perturb(cln_data, true_label)
-
-target = torch.ones_like(true_label) * 3
-adversary.targeted = True
-adv_targeted = adversary.perturb(cln_data, target)
-```
-
-```python
-
 logits_cal = ...
 Y_cal = ...
 
@@ -77,48 +37,21 @@ Y_Sets = predictor.predict(logits_test)
 
 # evaluate the prediction sets
 metrics = utils.coverage_rate(Y_sets,Y_test)
-
 ```
 
-For runnable examples see [`advertorch_examples/tutorial_attack_defense_bpda_mnist.ipynb`](https://github.com/BorealisAI/advertorch/blob/master/advertorch_examples/tutorial_attack_defense_bpda_mnist.ipynb) for how to attack and defend; see [`advertorch_examples/tutorial_train_mnist.py`](https://github.com/BorealisAI/advertorch/blob/master/advertorch_examples/tutorial_train_mnist.py) for how to adversarially train a robust model on MNIST.
-
-## Documentation
-
-The documentation webpage is on readthedocs  https://advertorch.readthedocs.io.
-
 
 ## Coming Soon
 
-AdverTorch is still under active development. We will add the following features/items down the road:
+DeepCP is still under active development. We will add the following features/items down the road:
 
-* more examples
-* support for other machine learning frameworks, e.g. TensorFlow
-* more attacks, defenses and other related functionalities
-* support for other Python versions and future PyTorch versions
-* contributing guidelines
+* more CP algorithms 
+* loss functions for CP
 * ...
 
-
-## Known issues
-
-`FastFeatureAttack` and `JacobianSaliencyMapAttack` do not pass the tests against the version of CleverHans used. (They use to pass tests on a previous version of CleverHans.) This issue is being investigated. In the file `test_attacks_on_cleverhans.py`, they are marked as "skipped" in `pytest` tests. 
-
 ## License
 
 This project is licensed under the LGPL. The terms and conditions can be found in the LICENSE and LICENSE.GPL files.
 
-## Citation
-
-If you use AdverTorch in your research, we kindly ask that you cite the following [technical report](https://arxiv.org/abs/1902.07623):
-
-```
-@article{ding2019advertorch,
-  title={{AdverTorch} v0.1: An Adversarial Robustness Toolbox based on PyTorch},
-  author={Ding, Gavin Weiguang and Wang, Luyu and Jin, Xiaomeng},
-  journal={arXiv preprint arXiv:1902.07623},
-  year={2019}
-}
-```
 
 
 ## Contributors

diff --git a/assets/advertorch.png b/assets/advertorch.png
diff --git a/assets/logo.png b/assets/logo.png
diff --git a/deepcp/VERSION b/deepcp/VERSION
diff --git a/deepcp/__init__.py b/deepcp/__init__.py
diff --git a/deepcp/attacks/__init__.py b/deepcp/attacks/__init__.py
diff --git a/deepcp/attacks/base.py b/deepcp/attacks/base.py
diff --git a/deepcp/attacks/blackbox/__init__.py b/deepcp/attacks/blackbox/__init__.py