Skip to content

Commit

Permalink
Add FedBPT research example (#2465)
Browse files Browse the repository at this point in the history 
* Add FedBPT research example

initial fedbpt files

add roberta model and run FL

move send to end

upgrade to 2.4.1rc and run experiment with 10 clients

move init to top

debug using pickle

record successful setting

use custom decomposer

clean code

add summary writer

add result figure

formatting

fix broken links

remove debug messages

update readme with system resources

use decomposer widget on server

* address comments; enable selection of evaluation client

* use new FedAvg api

* exclude dir from license test

* only exclude file for license check
  • Loading branch information
@holgerroth
holgerroth authored Apr 12, 2024
1 parent ef025f1 commit 68d42d8
Show file tree
Hide file tree
Showing 25 changed files with 5,242 additions and 6 deletions.
2 changes: 1 addition & 1 deletion integration/nemo/examples/peft/peft.ipynb
View file
Original file line number Diff line number Diff line change
Expand Up @@ -177,7 +177,7 @@
"metadata": {},
"source": [
"## Federated learning simulations\n",
"Next, we are using NVFlare's [simulator](https://nvflare.readthedocs.io/en/latest/user_guide/fl_simulator.html) to simulate each client training on their own dataset locally and all three clients training together using the [FedAvg](https://arxiv.org/abs/1602.05629) algorithm implemented in NVFlare.\n",
"Next, we are using NVFlare's [simulator](https://nvflare.readthedocs.io/en/latest/user_guide/nvflare_cli/fl_simulator.html) to simulate each client training on their own dataset locally and all three clients training together using the [FedAvg](https://arxiv.org/abs/1602.05629) algorithm implemented in NVFlare.\n",
"\n",
"With this setting, we require a GPU with at least 24GB of memory to run all clients in parallel on the same GPU. \n",
"If you have multiple GPUs in your system, you can use the `gpu` argument to assign one GPU for each client, e.g., `gpu=\"0,1\"`.\n",
Expand Down
2 changes: 1 addition & 1 deletion integration/nemo/examples/prompt_learning/prompt_learning.ipynb
View file
Original file line number Diff line number Diff line change
Expand Up @@ -171,7 +171,7 @@
"metadata": {},
"source": [
"## Federated learning simulations\n",
"Next, we are using NVFlare's [simulator](https://nvflare.readthedocs.io/en/latest/user_guide/fl_simulator.html) to simulate each client training on their own dataset locally and all three clients training together using the [FedAvg](https://arxiv.org/abs/1602.05629) algorithm implemented in NVFlare.\n",
"Next, we are using NVFlare's [simulator](https://nvflare.readthedocs.io/en/latest/user_guide/nvflare_cli/fl_simulator.html) to simulate each client training on their own dataset locally and all three clients training together using the [FedAvg](https://arxiv.org/abs/1602.05629) algorithm implemented in NVFlare.\n",
"\n",
"With this setting, we require a GPU with at least 16GB memory to run all clients in parallel on the same GPU. \n",
"If you have multiple GPUs in your system, you can use the `gpu` argument to assign one GPU for each client, e.g., `gpu=\"0,1\"`."
Expand Down
2 changes: 1 addition & 1 deletion research/auto-fed-rl/README.md
View file
Original file line number Diff line number Diff line change
Expand Up @@ -34,7 +34,7 @@ python3 -m pt.utils.cifar10_data_utils

## 3. Run simulated FL experiments

We are using NVFlare's [FL simulator](https://nvflare.readthedocs.io/en/latest/user_guide/fl_simulator.html) to run the following experiments.
We are using NVFlare's [FL simulator](https://nvflare.readthedocs.io/en/latest/user_guide/nvflare_cli/fl_simulator.html) to run the following experiments.

The output root of where to save the results is set in [./run_simulator.sh](./run_simulator.sh) as `RESULT_ROOT=/tmp/nvflare/sim_cifar10`.

Expand Down
201 changes: 201 additions & 0 deletions research/fed-bpt/LICENSE.txt
View file
Original file line number Diff line number Diff line change
@@ -0,0 +1,201 @@
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.
79 changes: 79 additions & 0 deletions research/fed-bpt/README.md
View file
Original file line number Diff line number Diff line change
@@ -0,0 +1,79 @@
# FedBPT: Efficient Federated Black-box Prompt Tuning for Large Language Models

This example shows how to run [FedBPT](https://arxiv.org/abs/2310.01467) on an example task and the [FL simulator](https://nvflare.readthedocs.io/en/latest/user_guide/nvflare_cli/fl_simulator.html).

###### Abstract:
> Pre-trained language models (PLM) have revolutionized the NLP landscape, achieving stellar performances across diverse tasks. These models, while benefiting from vast training data, often require fine-tuning on specific data to cater to distinct downstream tasks. However, this data adaptation process has inherent security and privacy concerns, primarily when leveraging user-generated, device-residing data. Federated learning (FL) provides a solution, allowing collaborative model fine-tuning without centralized data collection. However, applying FL to finetune PLMs is hampered by challenges, including restricted model parameter access, high computational requirements, and communication overheads. This paper introduces Federated Black-box Prompt Tuning (FedBPT), a framework designed to address these challenges. FedBPT does not require the clients to access the model parameters. By focusing on training optimal prompts and utilizing gradient-free optimization methods, FedBPT reduces the number of exchanged variables, boosts communication efficiency, and minimizes computational and storage costs. Experiments highlight the framework's ability to drastically cut communication and memory costs while maintaining competitive performance. Ultimately, FedBPT presents a promising solution for efficient, privacy-preserving fine-tuning of PLM in the age of large language models.

## License
The code in this directory is released under Apache v2 License.
The code is extended from [Black-Box-Tuning (BBT)](https://github.com/txsun1997/Black-Box-Tuning) which is released under MIT License.
The models code is copied from the [transformers](https://github.com/huggingface/transformers) library.

## 1. Setup
We recommend creating a [conda environment](https://www.anaconda.com) following [BBT](https://github.com/txsun1997/Black-Box-Tuning#prepare-your-environment)
with the addition of installing NVFlare for running federated learning and some other updates:
```commandline
conda create --name fedbpt python=3.8
conda activate fedbpt
pip install -r requirements.txt
```

## 2. Run a federated learning experiment
First, we set the location of NVFlare job templates directory.
```commandline
nvflare config -jt ./job_templates
```
Next, we generate a job configuration from the template to run FL on `N_CLIENTS` clients.
We utilize the [SST-2 dataset](https://huggingface.co/datasets/stanfordnlp/sst2) and the RoBerTa-large model for training.
```commandline
N_CLIENTS=10
SEED=1234
nvflare job create -force -j "./jobs/fedbpt" -w "fedbpt" -sd "./src" \
-f app/config/config_fed_client.conf app_script="fedbpt_train.py" app_config="--task_name sst2 \
--n_prompt_tokens 50 \
--intrinsic_dim 500 \
--k_shot 200 \
--device cuda:0 \
--seed ${SEED} \
--loss_type ce \
--cat_or_add add \
--local_iter 8 \
--num_users ${N_CLIENTS} \
--iid 1 \
--local_popsize 5 \
--perturb 1 \
--model_name roberta-large \
--eval_clients site-1 \
--llama_causal 1" \
-f app/config/config_fed_server.conf min_clients=${N_CLIENTS} num_rounds=200 seed=${SEED}
```
By default, we only evaluate the global model on client `site-1` as in our setting, the global test set is shared by clients.

Start the FL simulator with `N_CLIENTS` clients in parallel.
The following setting requires a GPU with at least 24 GB memory and enough system memory to run the clients in parallel (we recommend at least 40 GB).
For a system with less resources, you can set -t to be a lower number and simulate the clients running sequentially.
```commandline
OUT_DIR="/tmp/nvflare/fedbpt"
nvflare simulator ./jobs/fedbpt -n ${N_CLIENTS} -t ${N_CLIENTS} -w ${OUT_DIR}
```
If you have more GPUs available on your system, you can use the `--gpu` argument of the simulator to run clients on different GPUs in parallel.

## 3. Example results
The training results showing the global testing accuracy over 200 rounds is shown below.
The global learnt prompt using FedBPT achieves an accuracy of 0.8761 on the SST-2 test set.
<img src="./figs/global_test_acc.png" alt="FedBPT results" width="600"/>

## Citation

> Sun, Jingwei, et al. "FedBPT: Efficient Federated Black-box Prompt Tuning for Large Language Models." arXiv preprint arXiv:2310.01467 (2023).
BibTeX
```
@article{sun2023fedbpt,
title={FedBPT: Efficient Federated Black-box Prompt Tuning for Large Language Models},
author={Sun, Jingwei and Xu, Ziyue and Yin, Hongxu and Yang, Dong and Xu, Daguang and Chen, Yiran and Roth, Holger R},
journal={arXiv preprint arXiv:2310.01467},
year={2023}
}
```
Binary file added research/fed-bpt/figs/global_test_acc.png
View file
Loading
Sorry, something went wrong. Reload?
Sorry, we cannot display this file.
Sorry, this file is invalid so it cannot be displayed.
16 changes: 16 additions & 0 deletions research/fed-bpt/index.html
View file
Original file line number Diff line number Diff line change
@@ -0,0 +1,16 @@
<html xmlns="http://www.w3.org/1999/xhtml">
<head>
<title>FedBPT</title>
<meta http-equiv="refresh" content="0;URL='https://github.com/NVIDIA/NVFlare/tree/main/research/fed-bpt'" />
</head>
<body>
<p>
Redirecting to the project codebase:
</p>
<p>
<a href="https://github.com/NVIDIA/NVFlare/tree/main/research/fed-bpt">
https://github.com/NVIDIA/NVFlare/tree/main/research/fed-bpt
</a>
</p>
</body>
</html>
Loading

0 comments on commit 68d42d8

Please sign in to comment.