Initialize MTLoRA

.gitignore

@@ -0,0 +1,146 @@
+# Byte-compiled / optimized / DLL files
+__pycache__/
+*.py[cod]
+*$py.class
+
+# launch bash
+*.sh
+# nsight system report files
+*.nsys-rep
+*.sqlite
+
+# C extensions
+*.so
+
+# Distribution / packaging
+.Python
+build/
+develop-eggs/
+dist/
+downloads/
+eggs/
+.eggs/
+lib/
+lib64/
+parts/
+sdist/
+var/
+wheels/
+pip-wheel-metadata/
+share/python-wheels/
+*.egg-info/
+.installed.cfg
+*.egg
+MANIFEST
+
+# PyInstaller
+#  Usually these files are written by a python script from a template
+#  before PyInstaller builds the exe, so as to inject date/other infos into it.
+*.manifest
+*.spec
+
+# Installer logs
+pip-log.txt
+pip-delete-this-directory.txt
+
+# Unit test / coverage reports
+htmlcov/
+.tox/
+.nox/
+.coverage
+.coverage.*
+.cache
+nosetests.xml
+coverage.xml
+*.cover
+*.py,cover
+.hypothesis/
+.pytest_cache/
+
+# Translations
+*.mo
+*.pot
+
+# Django stuff:
+*.log
+local_settings.py
+db.sqlite3
+db.sqlite3-journal
+
+# Flask stuff:
+instance/
+.webassets-cache
+
+# Scrapy stuff:
+.scrapy
+
+# Sphinx documentation
+docs/_build/
+
+# PyBuilder
+target/
+
+# Jupyter Notebook
+.ipynb_checkpoints
+
+# IPython
+profile_default/
+ipython_config.py
+
+# pyenv
+.python-version
+
+# pipenv
+#   According to pypa/pipenv#598, it is recommended to include Pipfile.lock in version control.
+#   However, in case of collaboration, if having platform-specific dependencies or dependencies
+#   having no cross-platform support, pipenv may install dependencies that don't work, or not
+#   install all needed dependencies.
+#Pipfile.lock
+
+# PEP 582; used by e.g. github.com/David-OConnor/pyflow
+__pypackages__/
+
+# Celery stuff
+celerybeat-schedule
+celerybeat.pid
+
+# SageMath parsed files
+*.sage.py
+
+# Environments
+.env
+.venv
+env/
+venv/
+ENV/
+env.bak/
+venv.bak/
+
+# Spyder project settings
+.spyderproject
+.spyproject
+
+# Rope project settings
+.ropeproject
+
+# mkdocs documentation
+/site
+
+# mypy
+.mypy_cache/
+.dmypy.json
+dmypy.json
+
+# Pyre type checker
+.pyre/
+output/
+
+*.out
+*.log
+*.csv
+*.png
+*.pth
+*.pt
+*.ptl
+*.dec
+wandb/*

LICENSE

@@ -0,0 +1,21 @@
+    MIT License
+
+    Copyright (c) SCALE Lab, Brown University.
+
+    Permission is hereby granted, free of charge, to any person obtaining a copy
+    of this software and associated documentation files (the "Software"), to deal
+    in the Software without restriction, including without limitation the rights
+    to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+    copies of the Software, and to permit persons to whom the Software is
+    furnished to do so, subject to the following conditions:
+
+    The above copyright notice and this permission notice shall be included in all
+    copies or substantial portions of the Software.
+
+    THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+    IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+    FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+    AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+    LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+    OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+    SOFTWARE

README.md

@@ -0,0 +1,53 @@
+# MTLoRA: A Low-Rank Adaptation Approach for Efficient Multi-Task Learning
+
+## Introduction
+
+This is the official implementation of the paper: **MTLoRA: A Low-Rank Adaptation Approach for Efficient Multi-Task Learning** developed at [Brown University SCALE lab](https://scale-lab.github.io).
+
+This repository provides a Python-based implementation of MTLoRA including [`MTLoRALinear`](models/lora.py) (the main module) and MTL architectures. 
+
+The repository is built on top of [Swin-Transformer](https://github.com/microsoft/Swin-Transformer) and uses some modules from [Multi-Task-Learning-PyTorch](https://github.com/SimonVandenhende/Multi-Task-Learning-PyTorch).
+
+
+## How to Run
+
+Running MTLoRA code, is very simmilar to Swin's codebase:
+
+1. **Clone the repository**
+    ```bash
+    git clone https://github.com/scale-lab/MTLoRA.git
+    cd MTLoRA
+    ```
+
+2. **Install the prerequisites**
+    - Install `PyTorch>=1.12.0` and `torchvision>=0.13.0` with `CUDA>=11.6`
+    - Install dependencies: `pip install -r requirements.txt`
+
+3. **Run the code**
+        ```
+        python main.py --cfg configs/mtlora/tiny_448/<config>.yaml --pascal <path to pascal database> --tasks semseg,normals,sal,human_parts --batch-size <batch size> --ckpt-freq=20 --epoch=<num epochs> --resume-backbone <path to the weights of the chosen Swin variant>
+        ```
+
+  Swin variants and their weights can be found at the official [Swin Transformer repository](https://github.com/microsoft/Swin-Transformer). 
+
+  The outputs will be saved in `output/` folder unless overridden by the argument `--output`.
+
+## Authorship
+Since the release commit is squashed, the GitHub contributors tab doesn't reflect the authors' contributions. The following authors contributed equally to this codebase:
+- [Ahmed Agiza](https://github.com/ahmed-agiza)
+- [Marina Neseem](https://github.com/marina-neseem)
+
+## Citation
+If you find MTLoRA helpful in your research, please cite our paper:
+```
+@inproceedings{agiza2023mtlora,
+  title={MTLoRA: A Low-Rank Adaptation Approach for Efficient Multi-Task Learning},
+  author={Agiza, Ahmeda and Neseem, Marina and Reda, Sherief},
+  booktitle={},
+  pages={},
+  year={2023}
+}
+```
+
+## License
+MIT License. See [LICENSE](LICENSE) file