Add FedPer (#2266)

Co-authored-by: jafermarq <[email protected]>

baselines/fedper/LICENSE

@@ -0,0 +1,202 @@
+
+                                 Apache License
+                           Version 2.0, January 2004
+                        http://www.apache.org/licenses/
+
+   TERMS AND CONDITIONS FOR USE, REPRODUCTION, AND DISTRIBUTION
+
+   1. Definitions.
+
+      "License" shall mean the terms and conditions for use, reproduction,
+      and distribution as defined by Sections 1 through 9 of this document.
+
+      "Licensor" shall mean the copyright owner or entity authorized by
+      the copyright owner that is granting the License.
+
+      "Legal Entity" shall mean the union of the acting entity and all
+      other entities that control, are controlled by, or are under common
+      control with that entity. For the purposes of this definition,
+      "control" means (i) the power, direct or indirect, to cause the
+      direction or management of such entity, whether by contract or
+      otherwise, or (ii) ownership of fifty percent (50%) or more of the
+      outstanding shares, or (iii) beneficial ownership of such entity.
+
+      "You" (or "Your") shall mean an individual or Legal Entity
+      exercising permissions granted by this License.
+
+      "Source" form shall mean the preferred form for making modifications,
+      including but not limited to software source code, documentation
+      source, and configuration files.
+
+      "Object" form shall mean any form resulting from mechanical
+      transformation or translation of a Source form, including but
+      not limited to compiled object code, generated documentation,
+      and conversions to other media types.
+
+      "Work" shall mean the work of authorship, whether in Source or
+      Object form, made available under the License, as indicated by a
+      copyright notice that is included in or attached to the work
+      (an example is provided in the Appendix below).
+
+      "Derivative Works" shall mean any work, whether in Source or Object
+      form, that is based on (or derived from) the Work and for which the
+      editorial revisions, annotations, elaborations, or other modifications
+      represent, as a whole, an original work of authorship. For the purposes
+      of this License, Derivative Works shall not include works that remain
+      separable from, or merely link (or bind by name) to the interfaces of,
+      the Work and Derivative Works thereof.
+
+      "Contribution" shall mean any work of authorship, including
+      the original version of the Work and any modifications or additions
+      to that Work or Derivative Works thereof, that is intentionally
+      submitted to Licensor for inclusion in the Work by the copyright owner
+      or by an individual or Legal Entity authorized to submit on behalf of
+      the copyright owner. For the purposes of this definition, "submitted"
+      means any form of electronic, verbal, or written communication sent
+      to the Licensor or its representatives, including but not limited to
+      communication on electronic mailing lists, source code control systems,
+      and issue tracking systems that are managed by, or on behalf of, the
+      Licensor for the purpose of discussing and improving the Work, but
+      excluding communication that is conspicuously marked or otherwise
+      designated in writing by the copyright owner as "Not a Contribution."
+
+      "Contributor" shall mean Licensor and any individual or Legal Entity
+      on behalf of whom a Contribution has been received by Licensor and
+      subsequently incorporated within the Work.
+
+   2. Grant of Copyright License. Subject to the terms and conditions of
+      this License, each Contributor hereby grants to You a perpetual,
+      worldwide, non-exclusive, no-charge, royalty-free, irrevocable
+      copyright license to reproduce, prepare Derivative Works of,
+      publicly display, publicly perform, sublicense, and distribute the
+      Work and such Derivative Works in Source or Object form.
+
+   3. Grant of Patent License. Subject to the terms and conditions of
+      this License, each Contributor hereby grants to You a perpetual,
+      worldwide, non-exclusive, no-charge, royalty-free, irrevocable
+      (except as stated in this section) patent license to make, have made,
+      use, offer to sell, sell, import, and otherwise transfer the Work,
+      where such license applies only to those patent claims licensable
+      by such Contributor that are necessarily infringed by their
+      Contribution(s) alone or by combination of their Contribution(s)
+      with the Work to which such Contribution(s) was submitted. If You
+      institute patent litigation against any entity (including a
+      cross-claim or counterclaim in a lawsuit) alleging that the Work
+      or a Contribution incorporated within the Work constitutes direct
+      or contributory patent infringement, then any patent licenses
+      granted to You under this License for that Work shall terminate
+      as of the date such litigation is filed.
+
+   4. Redistribution. You may reproduce and distribute copies of the
+      Work or Derivative Works thereof in any medium, with or without
+      modifications, and in Source or Object form, provided that You
+      meet the following conditions:
+
+      (a) You must give any other recipients of the Work or
+          Derivative Works a copy of this License; and
+
+      (b) You must cause any modified files to carry prominent notices
+          stating that You changed the files; and
+
+      (c) You must retain, in the Source form of any Derivative Works
+          that You distribute, all copyright, patent, trademark, and
+          attribution notices from the Source form of the Work,
+          excluding those notices that do not pertain to any part of
+          the Derivative Works; and
+
+      (d) If the Work includes a "NOTICE" text file as part of its
+          distribution, then any Derivative Works that You distribute must
+          include a readable copy of the attribution notices contained
+          within such NOTICE file, excluding those notices that do not
+          pertain to any part of the Derivative Works, in at least one
+          of the following places: within a NOTICE text file distributed
+          as part of the Derivative Works; within the Source form or
+          documentation, if provided along with the Derivative Works; or,
+          within a display generated by the Derivative Works, if and
+          wherever such third-party notices normally appear. The contents
+          of the NOTICE file are for informational purposes only and
+          do not modify the License. You may add Your own attribution
+          notices within Derivative Works that You distribute, alongside
+          or as an addendum to the NOTICE text from the Work, provided
+          that such additional attribution notices cannot be construed
+          as modifying the License.
+
+      You may add Your own copyright statement to Your modifications and
+      may provide additional or different license terms and conditions
+      for use, reproduction, or distribution of Your modifications, or
+      for any such Derivative Works as a whole, provided Your use,
+      reproduction, and distribution of the Work otherwise complies with
+      the conditions stated in this License.
+
+   5. Submission of Contributions. Unless You explicitly state otherwise,
+      any Contribution intentionally submitted for inclusion in the Work
+      by You to the Licensor shall be under the terms and conditions of
+      this License, without any additional terms or conditions.
+      Notwithstanding the above, nothing herein shall supersede or modify
+      the terms of any separate license agreement you may have executed
+      with Licensor regarding such Contributions.
+
+   6. Trademarks. This License does not grant permission to use the trade
+      names, trademarks, service marks, or product names of the Licensor,
+      except as required for reasonable and customary use in describing the
+      origin of the Work and reproducing the content of the NOTICE file.
+
+   7. Disclaimer of Warranty. Unless required by applicable law or
+      agreed to in writing, Licensor provides the Work (and each
+      Contributor provides its Contributions) on an "AS IS" BASIS,
+      WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or
+      implied, including, without limitation, any warranties or conditions
+      of TITLE, NON-INFRINGEMENT, MERCHANTABILITY, or FITNESS FOR A
+      PARTICULAR PURPOSE. You are solely responsible for determining the
+      appropriateness of using or redistributing the Work and assume any
+      risks associated with Your exercise of permissions under this License.
+
+   8. Limitation of Liability. In no event and under no legal theory,
+      whether in tort (including negligence), contract, or otherwise,
+      unless required by applicable law (such as deliberate and grossly
+      negligent acts) or agreed to in writing, shall any Contributor be
+      liable to You for damages, including any direct, indirect, special,
+      incidental, or consequential damages of any character arising as a
+      result of this License or out of the use or inability to use the
+      Work (including but not limited to damages for loss of goodwill,
+      work stoppage, computer failure or malfunction, or any and all
+      other commercial damages or losses), even if such Contributor
+      has been advised of the possibility of such damages.
+
+   9. Accepting Warranty or Additional Liability. While redistributing
+      the Work or Derivative Works thereof, You may choose to offer,
+      and charge a fee for, acceptance of support, warranty, indemnity,
+      or other liability obligations and/or rights consistent with this
+      License. However, in accepting such obligations, You may act only
+      on Your own behalf and on Your sole responsibility, not on behalf
+      of any other Contributor, and only if You agree to indemnify,
+      defend, and hold each Contributor harmless for any liability
+      incurred by, or claims asserted against, such Contributor by reason
+      of your accepting any such warranty or additional liability.
+
+   END OF TERMS AND CONDITIONS
+
+   APPENDIX: How to apply the Apache License to your work.
+
+      To apply the Apache License to your work, attach the following
+      boilerplate notice, with the fields enclosed by brackets "[]"
+      replaced with your own identifying information. (Don't include
+      the brackets!)  The text should be enclosed in the appropriate
+      comment syntax for the file format. We also recommend that a
+      file or class name and description of purpose be included on the
+      same "printed page" as the copyright notice for easier
+      identification within third-party archives.
+
+   Copyright [yyyy] [name of copyright owner]
+
+   Licensed under the Apache License, Version 2.0 (the "License");
+   you may not use this file except in compliance with the License.
+   You may obtain a copy of the License at
+
+       http://www.apache.org/licenses/LICENSE-2.0
+
+   Unless required by applicable law or agreed to in writing, software
+   distributed under the License is distributed on an "AS IS" BASIS,
+   WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+   See the License for the specific language governing permissions and
+   limitations under the License.

baselines/fedper/README.md

@@ -0,0 +1,152 @@
+---
+title: Federated Learning with Personalization Layers
+url: https://arxiv.org/abs/1912.00818
+labels: [system heterogeneity, image classification, personalization, horizontal data partition] 
+dataset: [CIFAR-10, FLICKR-AES] 
+---
+
+# Federated Learning with Personalization Layers
+
+> Note: If you use this baseline in your work, please remember to cite the original authors of the paper as well as the Flower paper.
+
+**Paper:** [arxiv.org/abs/1912.00818](https://arxiv.org/abs/1912.00818)
+
+**Authors:** Manoj Ghuhan Arivazhagan, Vinay Aggarwal, Aaditya Kumar Singh, and Sunav Choudhary
+
+**Abstract:** The emerging paradigm of federated learning strives to enable collaborative training of machine learning models on the network edge without centrally aggregating raw data and hence, improving data privacy. This sharply deviates from traditional machine learning and necessitates design of algorithms robust to various sources of heterogeneity. Specifically, statistical heterogeneity of data across user devices can severely degrade performance of standard federated averaging for traditional machine learning applications like personalization with deep learning. This paper proposes `FedPer`, a base + personalization layer approach for federated training of deep feed forward neural networks, which can combat the ill-effects of statistical heterogeneity. We demonstrate effectiveness of `FedPer` for non-identical data partitions of CIFAR datasets and on a personalized image aesthetics dataset from Flickr.
+
+## About this baseline
+
+**What’s implemented:** The code in this directory replicates the experiments in _Federated Learning with Personalization Layers_ (Arivazhagan et al., 2019) for CIFAR10 and FLICKR-AES datasets, which proposed the `FedPer` model. Specifically, it replicates the results found in figures 2, 4, 7, and 8 in their paper. __Note__ that there is typo in the caption of Figure 4 in the article, it should be CIFAR10 and __not__ CIFAR100. 
+
+**Datasets:** CIFAR10 from PyTorch's Torchvision and FLICKR-AES. FLICKR-AES was proposed as dataset in _Personalized Image Aesthetics_ (Ren et al., 2017) and can be downloaded using a link provided on thier [GitHub](https://github.com/alanspike/personalizedImageAesthetics). One must first download FLICKR-AES-001.zip (5.76GB), extract all inside and place in baseline/FedPer/datasets. To this location, also download the other 2 related files: (1) FLICKR-AES_image_labeled_by_each_worker.csv, and (2) FLICKR-AES_image_score.txt. Images are also scaled to 224x224 for both datasets. This is not explicitly stated in the paper but seems to be boosting performance. Also, for FLICKR dataset, it is stated in the paper that they use data from clients with more than 60 and less than 290 rated images. This amounts to circa 60 clients and we randomly select 30 out of these (as in paper). Therefore, the results might differ somewhat but only slighly. Since the pre-processing steps in the paper are somewhat obscure, the metric values in the plots below may differ slightly, but not the overall results and findings. 
+
+```bash
+# These steps are not needed if you are only interested in CIFAR-10
+
+# Create the `datasets` directory if it doesn't exist already
+mkdir datasets
+
+# move/copy the downloaded FLICKR-AES-001.zip file to `datasets/`
+
+# unzip dataset to a directory named `flickr`
+cd datasets
+unzip FLICKR-AES-001.zip -d flickr
+
+# then move the .csv files inside flickr
+mv FLICKR-AES_image_labeled_by_each_worker.csv flickr
+mv FLICKR-AES_image_score.txt flickr
+```
+
+**Hardware Setup:** Experiments have been carried out on GPU. 2 different computers managed to run experiments: 
+
+- GeForce RTX 3080 16GB
+- GeForce RTX 4090 24GB
+
+It's worth mentioning that GPU memory for each client is ~7.5GB. When training on powerful GPUs, one can reduce ratio of GPU needed for each client in the configuration setting to e.g. `num_gpus` to 0.33. 
+
+> NOTE: One experiment carried out using 1 GPU (RTX 4090) takes somehwere between 1-3h depending on dataset and model. Running ResNet34 compared to MobileNet-v1 takes approximately 10-15% longer.  
+
+**Contributors:** [William Lindskog](https://github.com/WilliamLindskog)
+
+
+## Experimental Setup
+
+**Task:** Image Classification
+
+**Model:** This directory implements 2 models:
+
+- ResNet34 which can be imported directly (after having installed the packages) from PyTorch, using `from torchvision.models import resnet34 
+- MobileNet-v1
+
+Please see how models are implemented using a so called model_manager and model_split class since FedPer uses head and base layers in a neural network. These classes are defined in the models.py file and thereafter called when building new models in the directory /implemented_models. Please, extend and add new models as you wish. 
+
+**Dataset:** CIFAR10, FLICKR-AES. CIFAR10 will be partitioned based on number of classes for data that each client shall recieve e.g. 4 allocated classes could be [1, 3, 5, 9]. FLICKR-AES is an unbalanced dataset, so there we only apply random sampling. 
+
+**Training Hyperparameters:** The hyperparameters can be found in conf/base.yaml file which is the configuration file for the main script. 
+
+| Description | Default Value |
+| ----------- | ----- |
+| num_clients | 10 |
+| clients per round | 10 |
+| number of rounds | 50 |
+| client resources | {'num_cpus': 4, 'num_gpus': 1 }|
+| learning_rate | 0.01 |
+| batch_size | 128 |
+| optimizer | SGD |
+| algorithm | fedavg|
+
+**Stateful Clients:**
+In this Baseline (FedPer), we must store the state of the local client head while aggregation of body parameters happen at the server. Flower is currently making this possible but for the time being, we reside to storing client _head_ state in a folder called client_states. We store the values after each fit and evaluate function carried out on each client, and call for the state before executing these funcitons. Moreover, the state of a unique client is accessed using the client ID. 
+
+> NOTE: This is a work-around so that the local head parameters are not reset before each fit and evaluate. Nevertheless, it can come to change with future releases. 
+
+
+## Environment Setup
+
+To construct the Python environment follow these steps:
+
+```bash
+# Set Python 3.10
+pyenv local 3.10.6
+# Tell poetry to use python 3.10
+poetry env use 3.10.6
+
+# Install the base Poetry environment
+poetry install
+
+# Activate the environment
+poetry shell
+```
+
+## Running the Experiments
+```bash
+python -m fedper.main # this will run using the default settings in the `conf/base.yaml` 
+
+# When running models for flickr dataset, it is important to keep batch size at 4 or lower since some clients (for reproducing experiment) will have very few examples of one class
+```
+
+While the config files contain a large number of settings, the ones below are the main ones you'd likely want to modify to .
+```bash
+algorithm: fedavg, fedper # these are currently supported
+server_device: 'cuda:0', 'cpu'
+dataset.name: 'cifar10', 'flickr'
+num_classes: 10, 5 # respectively 
+dataset.num_classes: 4, 8, 10 # for non-iid split assigning n num_classes to each client (these numbers for CIFAR10 experiments)
+model_name: mobile, resnet
+```
+
+To run multiple runs, one can also reside to `HYDRA`'s multirun option. 
+```bash
+# for CIFAR10
+python -m fedper.main --multirun --config_name cifar10 dataset.num_classes=4,8,10 model_name=resnet,mobile algorithm=fedper,fedavg model.num_head_layers=2,3
+
+# to repeat each run 5 times, one can also add
+python -m fedper.main --multirun --config_name cifar10 dataset.num_classes=4,8,10 model_name=resnet,mobile algorithm=fedper,fedavg model.num_head_layers=2,3 '+repeat_num=range(5)'
+```
+
+
+## Expected Results
+
+To reproduce figures make `fedper/run_figures.sh` executable and run it. By default all experiments will be run:
+
+```bash
+# Make fedper/run_figures.sh executable
+chmod u+x fedper/run_figures.sh
+# Run the script 
+bash fedper/run_figures.sh
+```
+
+Having run the `run_figures.sh`, the expected results should look something like this: 
+
+**MobileNet-v1 and ResNet-34 on CIFAR10**
+
+<img src="_static/mobile_plot_figure_2.png" width="400"/> <img src="_static/resnet_plot_figure_2.png" width="400"/>
+
+**MobileNet-v1 and ResNet-34 on CIFAR10 using varying size of head**
+
+<img src="_static/mobile_plot_figure_num_head.png" width="400"/> <img src="_static/resnet_plot_figure_num_head.png" width="400"/>
+
+**MobileNet-v1 and ResNet-34 on FLICKR-AES**
+
+<img src="_static/mobile_plot_figure_flickr.png" width="400"/> <img src="_static/resnet_plot_figure_flickr.png" width="400"/>

baselines/fedper/fedper/__init__.py

@@ -0,0 +1 @@
+"""Template baseline package."""