Metadata-Version: 2.4
Name: mmlearn
Version: 0.1.0b1
Summary: A modular framework for research on multimodal representation learning.
Author-email: Vector AI Engineering <ai_engineering@vectorinstitute.ai>
License-Expression: Apache-2.0
License-File: LICENSE
Requires-Python: <3.13,>=3.10
Requires-Dist: bottleneck<2.0.0,>=1.4.2
Requires-Dist: hydra-core<2.0.0,>=1.3.2
Requires-Dist: hydra-submitit-launcher<2.0.0,>=1.2.0
Requires-Dist: hydra-zen<0.14.0,>=0.13.0
Requires-Dist: lightning<3.0.0,>=2.5.0.post0
Requires-Dist: numpy<2.0.0,>=1.26.4
Requires-Dist: pandas<3.0.0,>=2.2.3; extra == 'performance'
Requires-Dist: pyarrow<18.0.0,>=17.0.0
Requires-Dist: torch<2.6.0,>=2.5.1
Requires-Dist: torchvision<0.21.0,>=0.20.1
Requires-Dist: transformers<5.0.0,>=4.48.3
Provides-Extra: audio
Requires-Dist: torchaudio<2.6.0,>=2.5.1; extra == 'audio'
Provides-Extra: peft
Requires-Dist: peft<0.15.0,>=0.14.0; extra == 'peft'
Provides-Extra: vision
Requires-Dist: opencv-python<5.0.0,>=4.11.0.86; extra == 'vision'
Requires-Dist: timm<2.0.0,>=1.0.14; extra == 'vision'
Description-Content-Type: text/markdown

# mmlearn

[![code checks](https://github.com/VectorInstitute/mmlearn/actions/workflows/code_checks.yml/badge.svg)](https://github.com/VectorInstitute/mmlearn/actions/workflows/code_checks.yml)
[![integration tests](https://github.com/VectorInstitute/mmlearn/actions/workflows/integration_tests.yml/badge.svg)](https://github.com/VectorInstitute/mmlearn/actions/workflows/integration_tests.yml)
[![license](https://img.shields.io/github/license/VectorInstitute/mmlearn.svg)](https://github.com/VectorInstitute/mmlearn/blob/main/LICENSE)

*mmlearn* aims at enabling the evaluation of existing multimodal representation learning methods, as well as facilitating
experimentation and research for new techniques.

## Quick Start

### Installation

#### Prerequisites

The library requires Python 3.10 or later. We recommend using a virtual environment to manage dependencies. You can create
a virtual environment using the following command:

```bash
python3 -m venv /path/to/new/virtual/environment
source /path/to/new/virtual/environment/bin/activate
```

#### Installing binaries

To install the pre-built binaries, run:

```bash
python3 -m pip install mmlearn
```

<details>
<summary><b>Installation Options</b></summary>
You can install optional dependencies to enable additional features. Use one or more of the pip extras listed below to
install the desired dependencies.

<table>
<tr>
<th style="text-align: left; width: 150px"> pip extra </th>
<th style="text-align: center"> Dependencies </th>
<th style="text-align: center"> Notes </th>
</tr>

<tr>
<td>
vision
</td>
<td>
"torchvision", "opencv-python", "timm"
</td>
<td>
Enables image processing and vision tasks.
</td>
</tr>

<tr>
<td>
audio
</td>
<td>
"torchaudio"
</td>
<td>
Enables audio processing and tasks.
</td>
</tr>

<tr>
<td>
peft
</td>
<td>
"peft"
</td>
<td>
Uses the <a href=https://huggingface.co/docs/peft/index>PEFT</a> library to enable parameter-efficient fine-tuning.
</td>
</tr>

</table>

For example, to install the library with the `vision` and `audio` extras, run:

```bash
python3 -m pip install mmlearn[vision,audio]
```

</details>

#### Building from source

To install the library from source, run:

```bash
git clone https://github.com/VectorInstitute/mmlearn.git
cd mmlearn
python3 -m pip install -e .
```

### Running Experiments

We use [Hydra](https://hydra.cc/docs/intro/) and [hydra-zen](https://mit-ll-responsible-ai.github.io/hydra-zen/) to manage configurations
in the library.

For new experiments, it is recommended to create a new directory to store the configuration files. The directory should
have an `__init__.py` file to make it a Python package and an `experiment` folder to store the experiment configuration files.
This format allows the use of `.yaml` configuration files as well as Python modules (using [structured configs](https://hydra.cc/docs/tutorials/structured_config/intro/) or [hydra-zen](https://mit-ll-responsible-ai.github.io/hydra-zen/)) to define the experiment configurations.

To run an experiment, use the following command:

```bash
mmlearn_run 'hydra.searchpath=[pkg://path.to.config.directory]' +experiment=<name_of_experiment_yaml_file> experiment=your_experiment_name
```

Hydra will compose the experiment configuration from all the configurations in the specified directory as well as all the
configurations in the `mmlearn` package. *Note the dot-separated path to the directory containing the experiment configuration
files.*
One can add a path to `hydra.searchpath` either as a package (`pkg://path.to.config.directory`) or as a file system
(`file://path/to/config/directory`). However, new configs in `mmlearn` are added to hydra's external store inside
`path/to/config/directory/__init__.py` which is only interpreted when the config directory is added as a package.
Hence, please refrain from using the `file://` notation.

Hydra also allows for overriding configuration parameters from the command line. To see the available options and other information, run:

```bash
mmlearn_run 'hydra.searchpath=[pkg://path.to.config.directory]' +experiment=<name_of_experiment_yaml_file> --help
```

By default, the `mmlearn_run` command will run the experiment locally. To run the experiment on a SLURM cluster, we use
the [submitit launcher](https://hydra.cc/docs/plugins/submitit_launcher/) plugin built into Hydra. The following is an example
of how to run an experiment on a SLURM cluster:

```bash
mmlearn_run --multirun \
    hydra.launcher.mem_per_cpu=5G \
    hydra.launcher.qos=your_qos \
    hydra.launcher.partition=your_partition \
    hydra.launcher.gres=gpu:4 \
    hydra.launcher.cpus_per_task=8 \
    hydra.launcher.tasks_per_node=4 \
    hydra.launcher.nodes=1 \
    hydra.launcher.stderr_to_stdout=true \
    hydra.launcher.timeout_min=720 \
    'hydra.searchpath=[pkg://path.to.my_project.configs]' \
    +experiment=my_experiment \
    experiment_name=my_experiment_name
```

This will submit a job to the SLURM cluster with the specified resources.

**Note**: After the job is submitted, it is okay to cancel the program with `Ctrl+C`. The job will continue running on
the cluster. You can also add `&` at the end of the command to run it in the background.

## Summary of Implemented Methods

<table>
<tr>
<th style="text-align: left; width: 250px"> Pretraining Methods </th>
<th style="text-align: center"> Notes </th>
</tr>
<tr>
<td>

Contrastive Pretraining
</td>
<td>
Uses the contrastive loss to align the representations from <i>N</i> modalities. Supports sharing of encoders, projection heads
or postprocessing modules (e.g. logit/temperature scaling) across modalities. Also supports multi-task learning with auxiliary
unimodal tasks applied to specific modalities.
</td>
</tr>
<tr>
<td>

I-JEPA
</td>
<td>
The <a href=https://arxiv.org/pdf/2301.08243>Image-based Joint-Embedding Predictive Architecture</a> (I-JEPA) is a unimodal non-generative
self-supervised learning method that predicts the <i>representations</i> of several target blocks of an image given a context block
from the same image. This task can be combined with the contrastive pretraining task to learn multimodal representations from
paired and unpaired data.
</td>
</tr>
<tr>
<th style="text-align: left; width: 250px"> Evaluation Methods </th>
<th style="text-align: center"> Notes </th>
</tr>
<tr>
<td>

Zero-shot Cross-modal Retrieval
</td>
<td>
Evaluates the quality of the learned representations in retrieving the <i>k</i> most similar examples from a different modality,
using recall@k metric. This is applicable to any number of pairs of modalities at once, depending on memory constraints.
</td>
</tr>
<tr>
<td>

Zero-shot Classification
</td>
<td>
Evaluates the ability of a pre-trained encoder-based multimodal model to predict classes that were not explicitly seen
during training. The new classes are given as text prompts, and the query modality can be any of the supported modalities.
Binary and multi-class classification tasks are supported.
</td>
</tr>
</table>

## Contributing

If you are interested in contributing to the library, please see [CONTRIBUTING.MD](CONTRIBUTING.MD). This file contains
many details around contributing to the code base, including are development practices, code checks, tests, and more.
