lightly-train 0.2.4__tar.gz → 0.3.0__tar.gz

{lightly_train-0.2.4 → lightly_train-0.3.0}/.github/pull_request_template.md

@@ -4,4 +4,4 @@
 
 ## How has it been tested?
 
-(Delete this: Please describe the tests that you ran to verify your changes. Provide instructions so we can reproduce. Please also list any relevant details for your test configuration.)
+(Delete this: Please describe the tests that you ran to verify your changes. Provide instructions so we can reproduce. Please also list any relevant details for your test configuration.)

{lightly_train-0.2.4 → lightly_train-0.3.0}/.github/workflows/build_docker_image.yml

@@ -1,4 +1,4 @@
-name: Docker image build and push to gcloud registry
+name: Build Docker Image
 
 # Pushes a release candidate build to lightly-main/train after a merge into main
 
@@ -18,7 +18,7 @@ concurrency:
 
 jobs:
   build-docker-image:
-    name: build docker image
+    name: Build Docker Image
     runs-on: cpu
     timeout-minutes: 60
     outputs:
@@ -41,7 +41,7 @@ jobs:
           cd ./docker && make main-deploy-amd64-cuda
 
   slack-notification:
-    name: send slack notification
+    name: Send Slack Notification
     needs:
       - build-docker-image
     if: always()

lightly_train-0.2.4/.github/workflows/test_code_format.yml → lightly_train-0.3.0/.github/workflows/check_code_format.yml

@@ -1,4 +1,4 @@
-name: Code Format Check
+name: Check Code Format
 
 on:
   push:
@@ -28,9 +28,9 @@ jobs:
               - '!docs/**'
               - '!docker/**'
               - '!.github/**'
-              - '.github/workflows/test_code_format.yml'
-  test:
-    name: Check
+              - '.github/workflows/check_code_format.yml'
+  check-code-format:
+    name: Check Code Format
     runs-on: cpu
     strategy:
       matrix:

{lightly_train-0.2.4 → lightly_train-0.3.0}/.github/workflows/release_dockerhub.yml

@@ -1,12 +1,12 @@
-name: Docker Hub Release
+name: Release DockerHub
 
 on:
   workflow_call:
   workflow_dispatch:
 
 jobs:
-  dockerhub-push:
-    name: Push the latest main image to DockerHub
+  release-dockerhub:
+    name: Release DockerHub
     runs-on: cpu
     defaults:
       run:
@@ -23,15 +23,15 @@ jobs:
         docker login -u borislightly -p ${{ secrets.DOCKERHUB_PASSWORD_BORISLIGHTLY }}
         make release
   slack-notification:
-    name: send slack notification
+    name: Send Slack Notification
     needs:
-      - dockerhub-push
+      - release-dockerhub
     if: always()
     runs-on: cpu
     timeout-minutes: 5
     steps:
     - name: Slack Notification - Success
-      if: ${{ needs.dockerhub-push.result == 'success' }}
+      if: ${{ needs.release-dockerhub.result == 'success' }}
       uses: rtCamp/action-slack-notify@v2
       env:
         SLACK_WEBHOOK: ${{ secrets.SLACK_WEBHOOK_RELEASES }}
@@ -44,12 +44,12 @@ jobs:
 
           Pull it from the registry with:
           ```
-          docker pull lighly/train:latest
+          docker pull lightly/train:latest
           ```
 
           See it in DockerHub at https://hub.docker.com/repository/docker/lightly/train/general
     - name: Slack Notification - Failure
-      if: ${{ needs.dockerhub-push.result != 'success' }}
+      if: ${{ needs.release-dockerhub.result != 'success' }}
       uses: rtCamp/action-slack-notify@v2
       env:
         SLACK_WEBHOOK: ${{ secrets.SLACK_WEBHOOK_RELEASES }}

{lightly_train-0.2.4 → lightly_train-0.3.0}/.github/workflows/release_pypi.yml

@@ -1,4 +1,4 @@
-name: Build and release LightlyTrain package to PyPI
+name: Release PyPI
 
 on:
   workflow_dispatch:
@@ -7,8 +7,8 @@ on:
 #    types: [published]
 
 jobs:
-  build:
-    name: Build and release
+  release-pypi:
+    name: Release PyPI
     runs-on: cpu
     steps:
     - name: Checkout code

{lightly_train-0.2.4 → lightly_train-0.3.0}/.github/workflows/test_docker.yml

@@ -1,4 +1,4 @@
-name: Docker Tests
+name: Test Docker
 
 on:
   push:
@@ -28,8 +28,8 @@ jobs:
               - '!docs/**'
               - '!.github/**'
               - '.github/workflows/test_docker.yml'
-  test:
-    name: Test
+  test-docker:
+    name: Test Docker
     needs: detect-code-changes
     if: needs.detect-code-changes.outputs.run-tests == 'true'
     runs-on: cpu

lightly_train-0.2.4/.github/workflows/test.yml → lightly_train-0.3.0/.github/workflows/test_unit.yml

@@ -1,4 +1,4 @@
-name: Unit Tests
+name: Test Unit
 
 on:
   push:
@@ -28,9 +28,9 @@ jobs:
               - '!docs/**'
               - '!docker/**'
               - '!.github/**'
-              - '.github/workflows/test.yml'
-  test:
-    name: Test
+              - '.github/workflows/test_unit.yml'
+  test-unit:
+    name: Test Unit
     needs: detect-code-changes
     if: needs.detect-code-changes.outputs.run-tests == 'true'
     runs-on: cpu
@@ -51,8 +51,6 @@ jobs:
           make install-pinned-${{ matrix.python }}
           make install-ffmpeg-ubuntu
       - name: Run Pytest
-        env:
-          CUDA_VISIBLE_DEVICES: ""
         run: |
           source .venv/bin/activate
           make test-ci

lightly_train-0.2.4/.github/workflows/weekly_dependency_tests.yml → lightly_train-0.3.0/.github/workflows/test_unit_latest_dependencies.yml

@@ -1,4 +1,4 @@
-name: Weekly Dependency Test
+name: Test Unit Latest Dependencies
 
 on:
   workflow_dispatch:
@@ -26,9 +26,9 @@ jobs:
               - '!docs/**'
               - '!docker/**'
               - '!.github/**'
-              - '.github/workflows/weekly_dependency_tests.yml'
-  test:
-    name: Test
+              - '.github/workflows/test_unit_latest_dependencies.yml'
+  test-unit-latest-dependencies:
+    name: Test Unit Latest Dependencies
     needs: detect-code-changes
     if: needs.detect-code-changes.outputs.run-tests == 'true'
     runs-on: cpu
@@ -49,6 +49,7 @@ jobs:
           make install-latest
           make install-ffmpeg-ubuntu
       - name: Run Pytest
+        # TODO(Philipp, 09/24): Remove CUDA_VISIBLE_DEVICES: "".
         env:
           CUDA_VISIBLE_DEVICES: ""
         run: |

lightly_train-0.2.4/.github/workflows/test_minimal_deps.yml → lightly_train-0.3.0/.github/workflows/test_unit_minimal_dependencies.yml

@@ -1,4 +1,4 @@
-name: Minimal Dependency Tests
+name: Test Unit Minimal Dependencies
 
 on:
   push:
@@ -28,9 +28,9 @@ jobs:
               - '!docs/**'
               - '!docker/**'
               - '!.github/**'
-              - '.github/workflows/test_minimal_deps.yml'
-  test:
-    name: Test
+              - '.github/workflows/test_unit_minimal_dependencies.yml'
+  test-unit-minimal-dependencies:
+    name: Test Unit Minimal Dependencies
     needs: detect-code-changes
     if: needs.detect-code-changes.outputs.run-tests == 'true'
     runs-on: cpu
@@ -54,8 +54,6 @@ jobs:
             make install-ffmpeg-ubuntu
           fi
       - name: Run Pytest
-        env:
-          CUDA_VISIBLE_DEVICES: ""
         run: |
           source .venv/bin/activate
           make test-ci

{lightly_train-0.2.4 → lightly_train-0.3.0}/.gitignore

@@ -2,6 +2,11 @@
 lightly_train_docker_test_data
 out
 
+### Ultralytics ###
+# Ultralytics downloads files to the current work directory.
+yolo*.pt
+yolo*.yaml
+
 # Created by https://www.toptal.com/developers/gitignore/api/macos,windows,linux,python
 # Edit at https://www.toptal.com/developers/gitignore?templates=macos,windows,linux,python
 

lightly_train-0.3.0/CONTRIBUTING.md

@@ -0,0 +1,48 @@
+# Contributing to LightlyTrain
+
+## Development
+
+```
+git clone https://github.com/lightly-ai/lightly-train.git
+make install-dev
+```
+
+```
+make format
+make static-checks
+make test
+```
+
+### Documentation
+
+Documentation is in the [docs](./docs) folder. To build the documentation, install
+dev dependencies with `make install-dev`, then move to the `docs` folder and run:
+
+```
+make docs
+```
+
+This builds the documentation in the `docs/build/<version>` folder.
+
+To build the documentation for the stable version, checkout the branch with the
+stable version and run:
+
+```
+make docs-stable
+```
+
+This builds the documentaion in the `docs/build/stable` folder.
+
+Docs can be served locally with:
+
+```
+make serve
+```
+
+#### Writing Documentation
+
+The documentation source is in [docs/source](./docs/source). The documentation is
+written in Markdown (MyST flavor). For more information regarding formatting, see:
+
+- https://pradyunsg.me/furo/reference/
+- https://myst-parser.readthedocs.io/en/latest/syntax/typography.html

{lightly_train-0.2.4 → lightly_train-0.3.0}/DOCS.md

@@ -9,6 +9,7 @@ pip install lightly-train[tensorboard,timm]
 ### Python
 
 #### Training
+
 ```python
 import lightly_train
 
@@ -28,20 +29,24 @@ lightly_train.train(
 In most cases you only have to specify `out`, `data`, and `model`. The rest is optional.
 
 You can monitor your training process with the help of `tensorboard` and `wandb` loggers:
+
 ```
 pip install "lightly-train[tensorboard, wandb]"
 ```
 
 Configure the loggers from Python:
-* `loggers={"tensorboard": True}`: Enable `tensorboard` logger with default arguments.
-* `loggers={"wandb": True}`: Enable `wandb` logger with default arguments.
-* `loggers={"wandb": {"project": "my-project"}}`: Configure `wandb` logger with custom arguments.
+
+- `loggers={"tensorboard": True}`: Enable `tensorboard` logger with default arguments.
+- `loggers={"wandb": True}`: Enable `wandb` logger with default arguments.
+- `loggers={"wandb": {"project": "my-project"}}`: Configure `wandb` logger with custom arguments.
 
 LightlyTrain uses the PyTorchLightning loggers under the hood. Learn more about their configuration:
-* https://lightning.ai/docs/pytorch/stable/api/lightning.pytorch.loggers.wandb.html
-* https://lightning.ai/docs/pytorch/stable/api/lightning.pytorch.loggers.tensorboard.html
+
+- https://lightning.ai/docs/pytorch/stable/api/lightning.pytorch.loggers.wandb.html
+- https://lightning.ai/docs/pytorch/stable/api/lightning.pytorch.loggers.tensorboard.html
 
 #### Exporting
+
 ```python
 import lightly_train
 
@@ -54,6 +59,7 @@ lightly_train.export(
 ```
 
 #### Embedding
+
 ```python
 import lightly_train
 
@@ -66,27 +72,29 @@ lightly_train.embed(
 ```
 
 #### Supported Models
+
 ```python
 import lightly_train
 print(lightly_train.list_models())
 ```
 
 #### Supported Methods
+
 ```python
 import lightly_train
 print(lightly_train.list_methods())
 ```
 
-
-
 ### Command Line
 
 #### Help
+
 ```
 lightly-train help
 ```
 
 #### Training
+
 ```
 lightly-train train \
     out=my_output_dir \
@@ -103,25 +111,28 @@ lightly-train train \
 In most cases you only have to specify `out`, `data`, and `model`. The rest is optional.
 
 You can monitor your training process with the help of `tensorboard` and `wandb` loggers:
+
 ```
 pip install "lightly-train[tensorboard, wandb]"
 ```
 
 Configure the loggers from the command-line:
-* `loggers.tensorboard=True`: Enable `tensorboard` logger with default arguments.
-* `loggers.wandb=True`: Enable `wandb` logger with default arguments.
-* `loggers.wandb.project="my-project"`: Configure `wandb` logger with custom arguments.
+
+- `loggers.tensorboard=True`: Enable `tensorboard` logger with default arguments.
+- `loggers.wandb=True`: Enable `wandb` logger with default arguments.
+- `loggers.wandb.project="my-project"`: Configure `wandb` logger with custom arguments.
 
 LightlyTrain uses the PyTorchLightning loggers under the hood. Learn more about their configuration:
-* https://lightning.ai/docs/pytorch/stable/api/lightning.pytorch.loggers.wandb.html
-* https://lightning.ai/docs/pytorch/stable/api/lightning.pytorch.loggers.tensorboard.html
 
+- https://lightning.ai/docs/pytorch/stable/api/lightning.pytorch.loggers.wandb.html
+- https://lightning.ai/docs/pytorch/stable/api/lightning.pytorch.loggers.tensorboard.html
 
 ```
 tensorboard --logdir my_output_dir
 ```
 
 #### Embedding
+
 ```
 lightly-train embed \
     out=my_output_dir/embeddings.csv \
@@ -131,6 +142,7 @@ lightly-train embed \
 ```
 
 #### Exporting
+
 ```
 lightly-train export \
     out=my_output_dir/model_state_dict.pt \
@@ -139,13 +151,14 @@ lightly-train export \
     format=torch_state_dict
 ```
 
-
 #### Supported Models
+
 ```
 lightly-train list_models
 ```
 
 #### Supported Methods
+
 ```
 lightly-train list_methods
 ```

lightly_train-0.3.0/MAINTAINING.md

@@ -0,0 +1,19 @@
+# Maintaining LightlyTrain
+
+This document is intended for maintainers of LightlyTrain. If you would like to
+contribute, please refer to the [CONTRIBUTING.md](CONTRIBUTING.md) document.
+
+## Release
+
+Follow these steps to release a new version of LightlyTrain:
+
+1. Bump the version in `lightly_train/__init__.py`. We use [Semantic Versioning](https://semver.org/). Note that we are still on major version 0 and therefore minor version
+   bumps are considered breaking changes.
+1. Once the version is bumped, create a new release on GitHub: https://github.com/lightly-ai/lightly-train/releases with a new tag that matches the version number. Make sure to
+   create release notes using the "Generate release notes" button.
+1. To release on PyPI, run the [PyPI Release action](https://github.com/lightly-ai/lightly-train/actions/workflows/release_pypi.yml) from the main branch.
+1. To release on DockerHub, run the [DockerHub Release action](https://github.com/lightly-ai/lightly-train/actions/workflows/release_dockerhub.yml) from the branch with the new
+   version tag.
+1. If the README was updated, make sure to copy the information to the DockerHub
+   description for LightlyTrain: https://hub.docker.com/r/lightly/ (you might have to ask
+   someone who is a member of the Lightly DockerHub organization to do this for you).

{lightly_train-0.2.4 → lightly_train-0.3.0}/Makefile

@@ -36,6 +36,11 @@ all-checks: static-checks test
 .PHONY: static-checks
 static-checks: format-check type-check
 
+# Files to format with mdformat.
+# This is needed to avoid formatting files in .venv. The mdformat command has an
+# --exclude option but only on Python 3.13+.
+MDFORMAT_FILES=.github docker docs src tests *.md
+
 # run formatter
 .PHONY: format
 format:
@@ -43,12 +48,16 @@ format:
 	ruff format .
 	# Fix linting issues and sort imports
 	ruff check --fix .
+	# Format markdown files
+	mdformat ${MDFORMAT_FILES}
 
 # run format check
 .PHONY: format-check
 format-check:
 	# Check code formatting
 	ruff check .
+	# Check markdown formatting
+	mdformat --check ${MDFORMAT_FILES}
 
 # run type check
 .PHONY: type-check
@@ -105,12 +114,19 @@ endif
 
 # SuperGradients is not compatible with Python>=3.10. It is also not easy to install
 # on MacOS. Therefore we exclude it from the default extras.
-EXTRAS = [dev,tensorboard,timm,wandb]
+EXTRAS = [dev,tensorboard,timm,ultralytics,wandb]
+# Exclude ultralytics from docker extras as it has an AGPL license and we should not
+# distribute it with the docker image.
 DOCKER_EXTRAS = --extra tensorboard --extra timm --extra wandb
 
 # Date until which dependencies installed with --exclude-newer must have been released.
 # Dependencies released after this date are ignored.
 EXCLUDE_NEWER_DATE="2024-08-28"
+# Pinned versions for Torch and TorchVision to avoid issues with the CUDA version on the CI machine.
+# These versions are compatible with CUDA 11.4 and Python 3.8. They are the latest versions
+# available before 2024-08-28.
+PINNED_TORCH_VERSION="2.4.0+cu118"
+PINNED_TORCHVISION_VERSION="0.19.0+cu118"
 
 # Install ffmpeg on Ubuntu.
 . PHONY: install-ffmpeg-ubuntu
@@ -157,9 +173,16 @@ install-minimal-extras:
 # We have to differentiate between Python versions because SuperGradients is not
 # compatible with Python>=3.10. For 3.8 we install the package with all extras,
 # including SuperGradients.
+#
+# Torch and TorchVision are pinned to specific versions to avoid issues with the
+# CUDA version on the CI machine. See PINNED_TORCH_VERSION and PINNED_TORCHVISION_VERSION
+# for the pinned versions.
 .PHONY: install-pinned-3.8
 install-pinned-3.8:
 	uv pip install --exclude-newer ${EXCLUDE_NEWER_DATE} --reinstall ${EDITABLE} . --all-extras --requirement pyproject.toml
+	uv pip install \
+		"torch==${PINNED_TORCH_VERSION}" "torchvision==${PINNED_TORCHVISION_VERSION}" \
+		--index-url https://download.pytorch.org/whl/cu118 --reinstall
 
 # Install package for Python 3.11 with dependencies pinned to the latest compatible
 # version available at EXCLUDE_NEWER_DATE.
@@ -168,6 +191,9 @@ install-pinned-3.8:
 .PHONY: install-pinned-3.11
 install-pinned-3.11:
 	uv pip install --exclude-newer ${EXCLUDE_NEWER_DATE} --reinstall ${EDITABLE} ".${EXTRAS}"
+	uv pip install \
+		"torch==${PINNED_TORCH_VERSION}" "torchvision==${PINNED_TORCHVISION_VERSION}" \
+		--index-url https://download.pytorch.org/whl/cu118 --reinstall
 
 # Install package with the latest dependencies.
 .PHONY: install-latest

{lightly_train-0.2.4/src/lightly_train.egg-info → lightly_train-0.3.0}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: lightly_train
-Version: 0.2.4
+Version: 0.3.0
 Summary: Train models with self-supervised learning in a single command
 Author: Lightly Team
 License: AGPL-3.0
@@ -8,26 +8,38 @@ Keywords: machine-learning,computer-vision,deep-learning,self-supervised-learnin
 Requires-Python: >=3.8
 Description-Content-Type: text/markdown
 License-File: LICENSE
+Requires-Dist: eval-type-backport>=0.2
 Requires-Dist: lightly>=1.5.12
 Requires-Dist: omegaconf>=2.3
 Requires-Dist: psutil>=5.0
+Requires-Dist: pyarrow>=15.0.0
 Requires-Dist: pytorch_lightning>=2.1
 Requires-Dist: torch>=2.0.0
 Requires-Dist: torchvision>=0.15.1
 Requires-Dist: tqdm>=4.0.0
 Provides-Extra: dev
-Requires-Dist: pytest>=8.0.0; extra == "dev"
+Requires-Dist: licenseheaders>=0.8.8; extra == "dev"
+Requires-Dist: mdformat>=0.7.17; extra == "dev"
+Requires-Dist: mypy>=1.10.0; extra == "dev"
 Requires-Dist: pytest-mock>=3.14.0; extra == "dev"
+Requires-Dist: pytest>=8.0.0; extra == "dev"
 Requires-Dist: ruff>=0.4.0; extra == "dev"
-Requires-Dist: mypy>=1.10.0; extra == "dev"
+Requires-Dist: types-pillow>=10.2.0; extra == "dev"
 Requires-Dist: types-tqdm>=4.0.0; extra == "dev"
-Requires-Dist: licenseheaders>=0.8.8; extra == "dev"
+Requires-Dist: furo>=2024.04.27; extra == "dev"
+Requires-Dist: myst_parser>=3.0; extra == "dev"
+Requires-Dist: sphinx-autoapi>=3.0; extra == "dev"
+Requires-Dist: sphinx-copybutton>=0.5.2; extra == "dev"
+Requires-Dist: sphinx-inline-tabs>=2023.4.21; extra == "dev"
+Requires-Dist: sphinx>=7.1; extra == "dev"
 Provides-Extra: super-gradients
 Requires-Dist: super-gradients>=3.7; extra == "super-gradients"
 Provides-Extra: tensorboard
 Requires-Dist: tensorboard>=2.10.0; extra == "tensorboard"
 Provides-Extra: timm
 Requires-Dist: timm>=1.0.3; extra == "timm"
+Provides-Extra: ultralytics
+Requires-Dist: ultralytics>=8.1.0; extra == "ultralytics"
 Provides-Extra: wandb
 Requires-Dist: wandb>=0.12.10; extra == "wandb"
 
@@ -35,13 +47,12 @@ Requires-Dist: wandb>=0.12.10; extra == "wandb"
 
 *Train models with self-supervised learning in a single command*
 
-
 ## Why LightlyTrain
 
 LightlyTrain uses self-supervised learning (SSL) to train models on large datasets
 without the need for labels. It provides simple Python, Command Line, and Docker
 interfaces to train models with popular SSL methods such as SimCLR or DINO. The trained
-models are ideal starting points for fine-tuning on downstream tasks such as image 
+models are ideal starting points for fine-tuning on downstream tasks such as image
 classification, object detection, and segmentation or for generating image embeddings.
 Models trained with LightlyTrain result in improved performance, faster convergence, and
 better generalization compared to models trained without SSL. Image embeddings created
@@ -51,19 +62,18 @@ and seamlessly extend to new classes due to the unsupervised nature of SSL.
 Lightly is the expert in SSL for computer vision and developed LightlyTrain to simplify
 model training for any task and dataset.
 
-
 ## Features
 
-* Train models on any image data without labels
-* Train models from popular libraries such as [torchvision](https://github.com/pytorch/vision), [TIMM](https://github.com/huggingface/pytorch-image-models), and [SuperGradients](https://github.com/Deci-AI/super-gradients)
-* Train custom models
-* No SSL expertise required
-* Automatic SSL method selection (soon!)
-* Python, Command Line, and Docker support
-* Multi-GPU and multi-node (soon!) support
-* Export models for fine-tuning or inference
-* Generate and export image embeddings
-* Monitor training progress with TensorBoard, Weights & Biases, Neptune, etc. (soon!)
+- Train models on any image data without labels
+- Train models from popular libraries such as [torchvision](https://github.com/pytorch/vision), [TIMM](https://github.com/huggingface/pytorch-image-models), [Ultralytics](https://github.com/ultralytics/ultralytics), and [SuperGradients](https://github.com/Deci-AI/super-gradients)
+- Train custom models
+- No SSL expertise required
+- Automatic SSL method selection (soon!)
+- Python, Command Line, and Docker support
+- Multi-GPU and multi-node (soon!) support
+- Export models for fine-tuning or inference
+- Generate and export image embeddings
+- Monitor training progress with TensorBoard, Weights & Biases, Neptune, etc. (soon!)
 
 ## License
 
@@ -72,6 +82,6 @@ at info@lightly.ai for more information.
 
 ## Contact
 
-* [**Email**](info@lightly.ai)  
-* [**Website**](https://www.lightly.ai/lightlytrain)  
-* [**Discord**](https://discord.gg/xvNJW94)
+- [**Email**](info@lightly.ai)
+- [**Website**](https://www.lightly.ai/lightlytrain)
+- [**Discord**](https://discord.gg/xvNJW94)

{lightly_train-0.2.4 → lightly_train-0.3.0}/README.md

@@ -2,13 +2,12 @@
 
 *Train models with self-supervised learning in a single command*
 
-
 ## Why LightlyTrain
 
 LightlyTrain uses self-supervised learning (SSL) to train models on large datasets
 without the need for labels. It provides simple Python, Command Line, and Docker
 interfaces to train models with popular SSL methods such as SimCLR or DINO. The trained
-models are ideal starting points for fine-tuning on downstream tasks such as image 
+models are ideal starting points for fine-tuning on downstream tasks such as image
 classification, object detection, and segmentation or for generating image embeddings.
 Models trained with LightlyTrain result in improved performance, faster convergence, and
 better generalization compared to models trained without SSL. Image embeddings created
@@ -18,19 +17,18 @@ and seamlessly extend to new classes due to the unsupervised nature of SSL.
 Lightly is the expert in SSL for computer vision and developed LightlyTrain to simplify
 model training for any task and dataset.
 
-
 ## Features
 
-* Train models on any image data without labels
-* Train models from popular libraries such as [torchvision](https://github.com/pytorch/vision), [TIMM](https://github.com/huggingface/pytorch-image-models), and [SuperGradients](https://github.com/Deci-AI/super-gradients)
-* Train custom models
-* No SSL expertise required
-* Automatic SSL method selection (soon!)
-* Python, Command Line, and Docker support
-* Multi-GPU and multi-node (soon!) support
-* Export models for fine-tuning or inference
-* Generate and export image embeddings
-* Monitor training progress with TensorBoard, Weights & Biases, Neptune, etc. (soon!)
+- Train models on any image data without labels
+- Train models from popular libraries such as [torchvision](https://github.com/pytorch/vision), [TIMM](https://github.com/huggingface/pytorch-image-models), [Ultralytics](https://github.com/ultralytics/ultralytics), and [SuperGradients](https://github.com/Deci-AI/super-gradients)
+- Train custom models
+- No SSL expertise required
+- Automatic SSL method selection (soon!)
+- Python, Command Line, and Docker support
+- Multi-GPU and multi-node (soon!) support
+- Export models for fine-tuning or inference
+- Generate and export image embeddings
+- Monitor training progress with TensorBoard, Weights & Biases, Neptune, etc. (soon!)
 
 ## License
 
@@ -39,6 +37,6 @@ at info@lightly.ai for more information.
 
 ## Contact
 
-* [**Email**](info@lightly.ai)  
-* [**Website**](https://www.lightly.ai/lightlytrain)  
-* [**Discord**](https://discord.gg/xvNJW94)
+- [**Email**](info@lightly.ai)
+- [**Website**](https://www.lightly.ai/lightlytrain)
+- [**Discord**](https://discord.gg/xvNJW94)