Multi-modal Emotion Recognition on Drawings Project - inzva AI Projects #7
Need packages to
pip install ftfy regex tqdm
pip install git+https://github.com/openai/CLIP.git
python run.py experiment=clip_emoreccom.yaml
Dataset folder should follow the structure below
├── ...
│ ├── emotic
│ | ├── ade20k
│ | ├── emodb_small
│ | ├── framesdb
│ | ├── mscoco
│ ├── Annotations
│ | ├── Annotations.mat
Pretrained Models and threshold for inference can be found at : https://drive.google.com/drive/folders/1e-JLA7V73CQD5pjTFCSWnKCmB0gCpV1D
pretrained_weights folder should follow the structure below
📦pretrained_weights
┣ 📂models
┃ ┣ 📜model_body1.pth
┃ ┣ 📜model_context1.pth
┃ ┗ 📜model_emotic1.pth
┗ 📂places
┃ ┣ 📜resnet18_places365.pth.tar
┃ ┗ 📜resnet18_state_dict.pth
A clean and scalable template to kickstart your deep learning project 🚀⚡🔥
Click on Use this template to initialize new repository.
Suggestions are always welcome!
This template tries to be as general as possible - you can easily delete any unwanted features from the pipeline or rewire the configuration, by modifying behavior in src/train.py.
Effective usage of this template requires learning of a couple of technologies: PyTorch, PyTorch Lightning and Hydra. Knowledge of some experiment logging framework like Weights&Biases, Neptune or MLFlow is also recommended.
Why you should use it: it allows you to rapidly iterate over new models/datasets and scale your projects from small single experiments to hyperparameter searches on computing clusters, without writing any boilerplate code. To my knowledge, it's one of the most convenient all-in-one technology stack for Deep Learning research. Good starting point for reproducing papers, kaggle competitions or small-team research projects. It's also a collection of best practices for efficient workflow and reproducibility.
Why you shouldn't use it: Lightning and Hydra are not yet mature, which means you might run into some bugs sooner or later. Also, even though Lightning is very flexible, it's not well suited for every possible deep learning task.
PyTorch Lightning is a lightweight PyTorch wrapper for high-performance AI research. It makes your code neatly organized and provides lots of useful features, like ability to run model on CPU, GPU, multi-GPU cluster and TPU.
Hydra is an open-source Python framework that simplifies the development of research and other complex applications. The key feature is the ability to dynamically create a hierarchical configuration by composition and override it through config files and the command line. It allows you to conveniently manage experiments and provides many useful plugins, like Optuna Sweeper for hyperparameter search, or Ray Launcher for running jobs on a cluster.
- Predefined Structure: clean and scalable so that work can easily be extended and replicated (see #Project Structure)
- Rapid Experimentation: thanks to automating pipeline with config files and hydra command line superpowers (see #Your Superpowers)
- Little Boilerplate: so pipeline can be easily modified (see src/train.py)
- Main Configuration: main config file specifies default training configuration (see #Main Project Configuration)
- Experiment Configurations: stored in a separate folder, they can be composed out of smaller configs, override chosen parameters or define everything from scratch (see #Experiment Configuration)
- Workflow: comes down to 4 simple steps (see #Workflow)
- Experiment Tracking: many logging frameworks can be easily integrated! (see #Experiment Tracking)
- Logs: all logs (checkpoints, data from loggers, chosen hparams, etc.) are stored in a convenient folder structure imposed by Hydra (see #Logs)
- Hyperparameter Search: made easier with Hydra built in plugins like Optuna Sweeper (see #Hyperparameter Search)
- Tests: unit tests and command based tests (see #Tests)
- Extra Features: optional utilities to make your life easier (see #Extra Features)
- Best Practices: a couple of recommended tools, practices and standards for efficient workflow and reproducibility (see #Best Practices)
The directory structure of new project looks like this:
├── bash <- Bash scripts
│ ├── setup_conda.sh <- Setup conda environment
│ └── schedule.sh <- Schedule execution of many runs
│
├── configs <- Hydra configuration files
│ ├── callbacks <- Callbacks configs
│ ├── datamodule <- Datamodule configs
│ ├── experiment <- Experiment configs
│ ├── hparams_search <- Hyperparameter search configs
│ ├── mode <- Running mode configs
│ ├── logger <- Logger configs
│ ├── model <- Model configs
│ ├── trainer <- Trainer configs
│ │
│ └── config.yaml <- Main project configuration file
│
├── data <- Project data
│
├── logs <- Logs generated by Hydra and PyTorch Lightning loggers
│
├── notebooks <- Jupyter notebooks. Naming convention is a number (for ordering),
│ the creator's initials, and a short `-` delimited description, e.g.
│ `1.0-jqp-initial-data-exploration.ipynb`.
│
├── tests <- Tests of any kind
│ ├── helpers <- A couple of testing utilities
│ ├── shell <- Shell/command based tests
│ └── unit <- Unit tests
│
├── src
│ ├── callbacks <- Lightning callbacks
│ ├── datamodules <- Lightning datamodules
│ ├── models <- Lightning models
│ ├── utils <- Utility scripts
│ │
│ └── train.py <- Training pipeline
│
├── run.py <- Run pipeline with chosen experiment configuration
│
├── .env.example <- Template of the file for storing private environment variables
├── .gitignore <- List of files/folders ignored by git
├── .pre-commit-config.yaml <- Configuration of automatic code formatting
├── setup.cfg <- Configurations of linters and pytest
├── requirements.txt <- File for installing python dependencies
└── README.md
# clone project
git clone https://github.com/ashleve/lightning-hydra-template
cd lightning-hydra-template
# [OPTIONAL] create conda environment
bash bash/setup_conda.sh
# install requirements
pip install -r requirements.txtTemplate contains example with MNIST classification.
When running python run.py you should see something like this:
(click to expand)
Override any config parameter from command line
Hydra allows you to easily overwrite any parameter defined in your config.
python run.py trainer.max_epochs=20 model.lr=1e-4You can also add new parameters with
+sign.
python run.py +model.new_param="uwu"Train on CPU, GPU, multi-GPU and TPU
PyTorch Lightning makes it easy to train your models on different hardware.
# train on CPU
python run.py trainer.gpus=0
# train on 1 GPU
python run.py trainer.gpus=1
# train on TPU
python run.py +trainer.tpu_cores=8
# train with DDP (Distributed Data Parallel) (4 GPUs)
python run.py trainer.gpus=4 +trainer.accelerator='ddp'
# train with DDP (Distributed Data Parallel) (8 GPUs, 2 nodes)
python run.py trainer.gpus=4 +trainer.num_nodes=2 +trainer.accelerator='ddp'Train with mixed precision
# train with mixed precision (Apex level O1)
python run.py trainer.gpus=1 +trainer.amp_backend="apex" +trainer.precision=16 \
+trainer.amp_level="O1"
# train with mixed precision (Apex level O2)
python run.py trainer.gpus=1 +trainer.amp_backend="apex" +trainer.precision=16 \
+trainer.amp_level="O2"Train model with any logger available in PyTorch Lightning, like Weights&Biases
PyTorch Lightning provides convenient integrations with most popular logging frameworks, like Tensorboard, Neptune or simple csv files. Read more here. Using wandb requires you to setup account first. After that just complete the config as below.
Click here to see example wandb dashboard generated with this template.
# set project and entity names in `configs/logger/wandb`
wandb:
project: "your_project_name"
entity: "your_wandb_team_name"# train model with Weights&Biases (link to wandb dashboard should appear in the terminal)
python run.py logger=wandbUse different running modes
# debug mode changes logging folder to `logs/debug/`
# also enables default trainer debugging options from `configs/trainer/debug.yaml`
# also sets level of all command line loggers to 'DEBUG'
python run.py mode=debug
# experiment mode changes logging folder to `logs/experiments/name_of_your_experiment/`
# name is also used by experiment loggers (like tensorboard or wandb)
python run.py mode=exp name='my_new_experiment_253'Train model with chosen experiment config
Experiment configurations are placed in configs/experiment/.
python run.py experiment=example_simpleAttach some callbacks to run
Callbacks can be used for things such as as model checkpointing, early stopping and many more.
Callbacks configurations are placed in configs/callbacks/.
python run.py callbacks=defaultUse different tricks available in Pytorch Lightning
PyTorch Lightning provides about 40+ useful trainer flags.
# gradient clipping may be enabled to avoid exploding gradients
python run.py +trainer.gradient_clip_val=0.5
# stochastic weight averaging can make your models generalize better
python run.py +trainer.stochastic_weight_avg=true
# run validation loop 4 times during a training epoch
python run.py +trainer.val_check_interval=0.25
# accumulate gradients
python run.py +trainer.accumulate_grad_batches=10
# terminate training after 12 hours
python run.py +trainer.max_time="00:12:00:00"Easily debug
# run 1 train, val and test loop, using only 1 batch
python run.py +trainer.fast_dev_run=true
# print full weight summary of all PyTorch modules
python run.py trainer.weights_summary="full"
# raise exception, if any of the parameters or the loss are NaN or +/-inf
python run.py +trainer.terminate_on_nan=true
# quick debug with all the previous options enabled
# (this is also enabled when using `python run.py mode=debug`)
python run.py trainer=debug
# print execution time profiling after training ends
python run.py +trainer.profiler="simple"
# try overfitting to 1 batch
python run.py +trainer.overfit_batches=1 trainer.max_epochs=20
# use only 20% of the data
python run.py +trainer.limit_train_batches=0.2 \
+trainer.limit_val_batches=0.2 +trainer.limit_test_batches=0.2
# log second gradient norm of the model
python run.py +trainer.track_grad_norm=2Resume training from checkpoint
Checkpoint can be either path or URL. Path should be absolute!
python run.py +trainer.resume_from_checkpoint="/absolute/path/to/ckpt/name.ckpt"
⚠️ Currently loading ckpt in Lightning doesn't resume logger experiment, but it will be supported in future Lightning release.
Create a sweep over hyperparameters
# this will run 6 experiments one after the other,
# each with different combination of batch_size and learning rate
python run.py -m datamodule.batch_size=32,64,128 model.lr=0.001,0.0005
⚠️ Currently sweeps aren't failure resistant (if one job crashes than the whole sweep crashes), but it will be supported in future Hydra release.
Create a sweep over hyperparameters with Optuna
Using Optuna Sweeper plugin doesn't require you to code any boilerplate into your pipeline, everything is defined in a single config file!
# this will run hyperparameter search defined in `configs/hparams_search/mnist_optuna.yaml`
# over chosen experiment config
python run.py -m hparams_search=mnist_optuna experiment=example_simpleExecute all experiments from folder
Hydra provides special syntax for controlling behavior of multiruns. Learn more here. The command below executes all experiments from folder configs/experiment/.
python run.py -m 'experiment=glob(*)'Execute sweep on a remote AWS cluster
This should be achievable with simple config using Ray AWS launcher for Hydra. Example is not yet implemented in this template.
Use Hydra tab completion
Hydra allows you to autocomplete config argument overrides in shell as you write them, by pressing
tabkey. Learn more here.
First you will need to install Nvidia Container Toolkit to enable GPU support.
The template Dockerfile is provided on branch dockerfiles. Copy it to the template root folder.
To build the container use:
docker build -t <project_name> .To mount the project to the container use:
docker run -v $(pwd):/workspace/project --gpus all -it --rm <project_name>Uncomment Apex in Dockerfile for mixed-precision support.
Have a question? Found a bug? Missing a specific feature? Ran into a problem? Feel free to file a new issue or PR with respective title and description. If you already found a solution to your problem, don't hesitate to share it. Suggestions for new best practices and tricks are always welcome!
- First, you should probably get familiar with PyTorch Lightning
- Next, go through Hydra quick start guide and basic Hydra tutorial
By design, every run is initialized by run.py file. All PyTorch Lightning modules are dynamically instantiated from module paths specified in config. Example model config:
_target_: src.models.mnist_model.MNISTLitModel
input_size: 784
lin1_size: 256
lin2_size: 256
lin3_size: 256
output_size: 10
lr: 0.001Using this config we can instantiate the object with the following line:
model = hydra.utils.instantiate(config.model)This allows you to easily iterate over new models!
Every time you create a new one, just specify its module path and parameters in appriopriate config file.
The whole pipeline managing the instantiation logic is placed in src/train.py.
Location: configs/config.yaml
Main project config contains default training configuration.
It determines how config is composed when simply executing command python run.py.
It also specifies everything that shouldn't be managed by experiment configurations.
Show main project configuration
# specify here default training configuration
defaults:
- trainer: default.yaml
- model: mnist_model.yaml
- datamodule: mnist_datamodule.yaml
- callbacks: default.yaml # set this to null if you don't want to use callbacks
- logger: null # set logger here or use command line (e.g. `python run.py logger=wandb`)
- mode: default.yaml
- experiment: null
- hparams_search: null
# path to original working directory
# hydra hijacks working directory by changing it to the current log directory,
# so it's useful to have this path as a special variable
# learn more here: https://hydra.cc/docs/next/tutorials/basic/running_your_app/working_directory
work_dir: ${hydra:runtime.cwd}
# path to folder with data
data_dir: ${work_dir}/data/
# pretty print config at the start of the run using Rich library
print_config: True
# disable python warnings if they annoy you
ignore_warnings: TrueLocation: configs/experiment
You should store all your experiment configurations in this folder.
Experiment configurations allow you to overwrite parameters from main project configuration.
Simple example
# to execute this experiment run:
# python run.py experiment=example_simple
defaults:
- override /trainer: default.yaml
- override /model: mnist_model.yaml
- override /datamodule: mnist_datamodule.yaml
- override /callbacks: default.yaml
- override /logger: null
# all parameters below will be merged with parameters from default configurations set above
# this allows you to overwrite only specified parameters
seed: 12345
trainer:
max_epochs: 10
gradient_clip_val: 0.5
model:
lin1_size: 128
lin2_size: 256
lin3_size: 64
lr: 0.005
datamodule:
train_val_test_split: [55_000, 5_000, 10_000]
batch_size: 64Show advanced example
# to execute this experiment run:
# python run.py experiment=example_full
defaults:
- override /trainer: null
- override /model: null
- override /datamodule: null
- override /callbacks: null
- override /logger: null
# we override default configurations with nulls to prevent them from loading at all
# instead we define all modules and their paths directly in this config,
# so everything is stored in one place
seed: 12345
trainer:
_target_: pytorch_lightning.Trainer
gpus: 0
min_epochs: 1
max_epochs: 10
gradient_clip_val: 0.5
model:
_target_: src.models.mnist_model.MNISTLitModel
lr: 0.001
weight_decay: 0.00005
input_size: 784
lin1_size: 256
lin2_size: 256
lin3_size: 128
output_size: 10
datamodule:
_target_: src.datamodules.mnist_datamodule.MNISTDataModule
data_dir: ${data_dir}
train_val_test_split: [55_000, 5_000, 10_000]
batch_size: 64
num_workers: 0
pin_memory: False
logger:
wandb:
_target_: pytorch_lightning.loggers.wandb.WandbLogger
project: "lightning-hydra-template"
tags: ["best_model", "uwu"]
notes: "Description of this model."- Write your PyTorch Lightning model (see mnist_model.py for example)
- Write your PyTorch Lightning datamodule (see mnist_datamodule.py for example)
- Write your experiment config, containing paths to your model and datamodule
- Run training with chosen experiment config:
python run.py experiment=experiment_name
Hydra creates new working directory for every executed run.
By default, logs have the following structure:
│
├── logs
│ ├── runs # Folder for logs generated from single runs
│ │ ├── 2021-02-15 # Date of executing run
│ │ │ ├── 16-50-49 # Hour of executing run
│ │ │ │ ├── .hydra # Hydra logs
│ │ │ │ ├── wandb # Weights&Biases logs
│ │ │ │ ├── checkpoints # Training checkpoints
│ │ │ │ └── ... # Any other thing saved during training
│ │ │ ├── ...
│ │ │ └── ...
│ │ ├── ...
│ │ └── ...
│ │
│ └── multiruns # Folder for logs generated from multiruns (sweeps)
│ ├── 2021-02-15_16-50-49 # Date and hour of executing sweep
│ │ ├── 0 # Job number
│ │ │ ├── .hydra # Hydra logs
│ │ │ ├── wandb # Weights&Biases logs
│ │ │ ├── checkpoints # Training checkpoints
│ │ │ └── ... # Any other thing saved during training
│ │ ├── 1
│ │ ├── 2
│ │ └── ...
│ ├── ...
│ └── ...
│
You can change this structure by modifying paths in hydra configuration.
PyTorch Lightning supports the most popular logging frameworks:
Weights&Biases · Neptune · Comet · MLFlow · Tensorboard
These tools help you keep track of hyperparameters and output metrics and allow you to compare and visualize results. To use one of them simply complete its configuration in configs/logger and run:
python run.py logger=logger_nameYou can use many of them at once (see configs/logger/many_loggers.yaml for example).
You can also write your own logger.
Lightning provides convenient method for logging custom metrics from inside LightningModule. Read the docs here or take a look at MNIST example.
Defining hyperparameter optimization is as easy as adding new config file to configs/hparams_search.
Show example
defaults:
- override /hydra/sweeper: optuna
# choose metric which will be optimized by Optuna
optimized_metric: "val/acc"
hydra:
# here we define Optuna hyperparameter search
# it optimizes for value returned from function with @hydra.main decorator
# learn more here: https://hydra.cc/docs/next/plugins/optuna_sweeper
sweeper:
_target_: hydra_plugins.hydra_optuna_sweeper.optuna_sweeper.OptunaSweeper
storage: null
study_name: null
n_jobs: 1
# 'minimize' or 'maximize' the objective
direction: maximize
# number of experiments that will be executed
n_trials: 20
# choose Optuna hyperparameter sampler
# learn more here: https://optuna.readthedocs.io/en/stable/reference/samplers.html
sampler:
_target_: optuna.samplers.TPESampler
seed: 12345
consider_prior: true
prior_weight: 1.0
consider_magic_clip: true
consider_endpoints: false
n_startup_trials: 10
n_ei_candidates: 24
multivariate: false
warn_independent_sampling: true
# define range of hyperparameters
search_space:
datamodule.batch_size:
type: categorical
choices: [32, 64, 128]
model.lr:
type: float
low: 0.0001
high: 0.2
model.lin1_size:
type: categorical
choices: [32, 64, 128, 256, 512]
model.lin2_size:
type: categorical
choices: [32, 64, 128, 256, 512]
model.lin3_size:
type: categorical
choices: [32, 64, 128, 256, 512]Next, you can execute it with: python run.py -m hparams_search=mnist_optuna
Using this approach doesn't require you to add any boilerplate into your pipeline, everything is defined in a single config file. You can use different optimization frameworks integrated with Hydra, like Optuna, Ax or Nevergrad. The optimization_results.yaml will be available under logs/multirun folder.
The following is example of loading model from checkpoint and running predictions.
Show inference example
from PIL import Image
from torchvision import transforms
from src.models.mnist_model import MNISTLitModel
def predict():
"""Example of inference with trained model.
It loads trained image classification model from checkpoint.
Then it loads example image and predicts its label.
"""
# ckpt can be also a URL!
CKPT_PATH = "last.ckpt"
# load model from checkpoint
# model __init__ parameters will be loaded from ckpt automatically
# you can also pass some parameter explicitly to override it
trained_model = MNISTLitModel.load_from_checkpoint(checkpoint_path=CKPT_PATH)
# print model hyperparameters
print(trained_model.hparams)
# switch to evaluation mode
trained_model.eval()
trained_model.freeze()
# load data
img = Image.open("data/example_img.png").convert("L") # convert to black and white
# img = Image.open("data/example_img.png").convert("RGB") # convert to RGB
# preprocess
mnist_transforms = transforms.Compose(
[
transforms.ToTensor(),
transforms.Resize((28, 28)),
transforms.Normalize((0.1307,), (0.3081,)),
]
)
img = mnist_transforms(img)
img = img.reshape((1, *img.size())) # reshape to form batch of size 1
# inference
output = trained_model(img)
print(output)
if __name__ == "__main__":
predict()Template comes with example tests implemented with pytest library.
To execute them simply run:
# run all tests
pytest
# run tests from specific file
pytest tests/shell/test_basic_commands.py
# run all tests except the ones marked as slow
pytest -k "not slow"I often find myself running into bugs that come out only in edge cases or on some specific hardware/environment. To speed up the development, I usually constantly execute tests that run a couple of quick 1 epoch experiments, like overfitting to 10 batches, training on 25% of data, etc. Those kind of tests don't check for any specific output - they exist to simply verify that executing some commands doesn't end up in throwing exceptions. You can find them implemented in tests/shell folder.
You can easily modify the commands in the scripts for your use case. If even 1 epoch is too much for your model, then you can make it run for a couple of batches instead (by using the right trainer flags).
Template contains example callbacks enabling better Weights&Biases integration, which you can use as a reference for writing your own callbacks (see wandb_callbacks.py).
To support reproducibility:
- WatchModel
- UploadCodeAsArtifact
- UploadCheckpointsAsArtifact
To provide examples of logging custom visualisations with callbacks only:
- LogConfusionMatrix
- LogF1PrecRecHeatmap
- LogImagePredictions
To try all of the callbacks at once, use:
python run.py logger=wandb callbacks=wandbTo see the result of all the callbacks attached, take a look at this experiment dashboard.
Lightning supports multiple ways of doing distributed training.
The most common one is DDP, which spawns separate process for each GPU and averages gradients between them. To learn about other approaches read lightning docs.
You can run DDP on mnist example with 4 GPUs like this:
python run.py trainer.gpus=4 +trainer.accelerator="ddp"To reproduce previous experiment, simply load its config from logs:
python run.py --config-path /logs/runs/.../.hydra/ --config-name config.yamlThe config.yaml from .hydra folder contains all overriden parameters and sections.
Use Miniconda for GPU environments
Use miniconda for your python environments (it's usually unnecessary to install full anaconda environment, miniconda should be enough).
It makes it easier to install some dependencies, like cudatoolkit for GPU support.
Example installation:
wget https://repo.anaconda.com/miniconda/Miniconda3-latest-Linux-x86_64.sh
bash Miniconda3-latest-Linux-x86_64.shCreate environment using bash script provided in the template:
bash bash/setup_conda.shUse automatic code formatting
Use pre-commit hooks to standardize code formatting of your project and save mental energy.
Simply install pre-commit package with:
pip install pre-commitNext, install hooks from .pre-commit-config.yaml:
pre-commit installAfter that your code will be automatically reformatted on every new commit.
Currently template contains configurations of black (python code formatting), isort (python import sorting), flake8 (python code analysis) and prettier (yaml formating).
To reformat all files in the project use command:
pre-commit run -aSet private environment variables in .env file
System specific variables (e.g. absolute paths to datasets) should not be under version control or it will result in conflict between different users. Your private keys also shouldn't be versioned since you don't want them to be leaked.
Template contains .env.example file, which serves as an example. Create a new file called .env (this name is excluded from version control in .gitignore).
You should use it for storing environment variables like this:
MY_VAR=/home/user/my_system_pathAll variables from .env are loaded in run.py automatically.
Hydra allows you to reference any env variable in .yaml configs like this:
path_to_data: ${oc.env:MY_VAR}Name metrics using '/' character
Depending on which logger you're using, it's often useful to define metric name with / character:
self.log("train/loss", loss)This way loggers will treat your metrics as belonging to different sections, which helps to get them organised in UI.
Use torchmetrics
Use official torchmetrics library to ensure proper calculation of metrics. This is especially important for multi-GPU training!
For example, instead of calculating accuracy by yourself, you should use the provided Accuracy class like this:
from torchmetrics.classification.accuracy import Accuracy
class LitModel(LightningModule):
def __init__(self)
self.train_acc = Accuracy()
self.val_acc = Accuracy()
def training_step(self, batch, batch_idx):
...
acc = self.train_acc(predictions, targets)
self.log("train/acc", acc)
...
def validation_step(self, batch, batch_idx):
...
acc = self.val_acc(predictions, targets)
self.log("val/acc", acc)
...Make sure to use different metric instance for each step to ensure proper value reduction over all GPU processes.
Torchmetrics provides metrics for most use cases, like F1 score or confusion matrix. Read documentation for more.
Follow PyTorch Lightning style guide
The style guide is available here.
-
Be explicit in your init. Try to define all the relevant defaults so that the user doesn’t have to guess. Provide type hints. This way your module is reusable across projects!
class LitModel(LightningModule): def __init__(self, layer_size: int = 256, lr: float = 0.001):
-
Preserve the recommended method order.
class LitModel(LightningModule): def __init__(): ... def forward(): ... def training_step(): ... def training_step_end(): ... def training_epoch_end(): ... def validation_step(): ... def validation_step_end(): ... def validation_epoch_end(): ... def test_step(): ... def test_step_end(): ... def test_epoch_end(): ... def configure_optimizers(): ... def any_extra_hook(): ...
Version control your data and models with DVC
Use DVC to version control big files, like your data or trained ML models.
To initialize the dvc repository:
dvc initTo start tracking a file or directory, use dvc add:
dvc add data/MNISTDVC stores information about the added file (or a directory) in a special .dvc file named data/MNIST.dvc, a small text file with a human-readable format. This file can be easily versioned like source code with Git, as a placeholder for the original data:
git add data/MNIST.dvc data/.gitignore
git commit -m "Add raw data"Support installing project as a package
It allows other people to easily use your modules in their own projects.
Change name of the src folder to your project name and add setup.py file:
from setuptools import find_packages, setup
setup(
name="src", # you should change "src" to your project name
version="0.0.0",
description="Describe Your Cool Project",
author="",
author_email="",
# replace with your own github project link
url="https://github.com/ashleve/lightning-hydra-template",
install_requires=["pytorch-lightning>=1.2.0", "hydra-core>=1.0.6"],
packages=find_packages(),
)Now your project can be installed from local files:
pip install -e .Or directly from git repository:
pip install git+git://github.com/YourGithubName/your-repo-name.git --upgradeSo any file can be easily imported into any other file like so:
from project_name.models.mnist_model import MNISTLitModel
from project_name.datamodules.mnist_datamodule import MNISTDataModuleAutomatic activation of virtual environment and tab completion when entering folder
Create a new file called .autoenv (this name is excluded from version control in .gitignore).
You can use it to automatically execute shell commands when entering folder.
To setup this automation for bash, execute the following line:
echo "autoenv() { if [ -x .autoenv ]; then source .autoenv ; echo '.autoenv executed' ; fi } ; cd() { builtin cd \"\$@\" ; autoenv ; } ; autoenv" >> ~/.bashrcNow you can add any commands to your .autoenv file, e.g. activation of virtual environment and hydra tab completion:
# activate conda environment
conda activate myenv
# activate hydra tab completion for bash
eval "$(python run.py -sc install=bash)"
# enable aliases for debugging
alias debug='python run.py mode=debug'
alias debug2='python run.py mode=debug trainer.fast_dev_run=false trainer.max_epochs=1 trainer.gpus=0'
alias debug3='python run.py mode=debug trainer.fast_dev_run=false trainer.max_epochs=1 trainer.gpus=1'
alias debug_wandb='python run.py mode=debug trainer.fast_dev_run=false trainer.max_epochs=1 trainer.gpus=1 logger=wandb logger.wandb.project=tests'(these commands will be executed whenever you're openning or switching terminal to folder containing .autoenv file)
Lastly add execution previliges to your .autoenv file:
chmod +x .autoenv
Explanation
The mentioned line appends your .bashrc file with 2 commands:
autoenv() { if [ -x .autoenv ]; then source .autoenv ; echo '.autoenv executed' ; fi }- this declares theautoenv()function, which executes.autoenvfile if it exists in current work dir and has execution previligiescd() { builtin cd \"\$@\" ; autoenv ; } ; autoenv- this extends behaviour ofcdcommand, to make it executeautoenv()function each time you change folder in terminal or open new terminal
Accessing datamodule attributes in model
The simplest way is to pass datamodule attribute directly to model on initialization:
datamodule = hydra.utils.instantiate(config.datamodule)
model = hydra.utils.instantiate(config.model, some_param=datamodule.some_param)This is not a robust solution, since it assumes all your datamodules have some_param attribute available (otherwise the run will crash).
A better solution is to add Omegaconf resolver to your datamodule:
from omegaconf import OmegaConf
# you can place this snippet in your datamodule __init__()
resolver_name = "datamodule"
OmegaConf.register_new_resolver(
resolver_name,
lambda name: getattr(self, name),
use_cache=False
)This way you can reference any datamodule attribute from your config like this:
# this will get 'datamodule.some_param' field
some_parameter: ${datamodule: some_param}When later accessing this field, say in your lightning model, it will get automatically resolved based on all resolvers that are registered. Remember not to access this field before datamodule is initialized. You also need to set resolve to false in print_config() in run.py method or it will throw errors!
utils.print_config(config, resolve=False)Inspirations
This template was inspired by: PyTorchLightning/deep-learninig-project-template, drivendata/cookiecutter-data-science, tchaton/lightning-hydra-seed, Erlemar/pytorch_tempest, lucmos/nn-template.
Useful repositories
- pytorch/hydra-torch - resources for configuring PyTorch classes with Hydra,
- romesco/hydra-lightning - resources for configuring PyTorch Lightning classes with Hydra
- lucmos/nn-template - similar template
- PyTorchLightning/lightning-transformers - official Lightning Transformers repo built with Hydra
This project is licensed under the MIT License.
MIT License
Copyright (c) 2021 ashleve
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.
DELETE EVERYTHING ABOVE FOR YOUR PROJECT
What it does
Install dependencies
# clone project
git clone https://github.com/YourGithubName/your-repo-name
cd your-repo-name
# [OPTIONAL] create conda environment
bash bash/setup_conda.sh
# install requirements
pip install -r requirements.txtTrain model with default configuration
# default
python run.py
# train on CPU
python run.py trainer.gpus=0
# train on GPU
python run.py trainer.gpus=1Train model with chosen experiment configuration from configs/experiment/
python run.py experiment=experiment_nameYou can override any parameter from command line like this
python run.py trainer.max_epochs=20 datamodule.batch_size=64