PyTorch Lightning + Hydra. A feature-rich template for rapid, scalable and reproducible ML experimentation with best practices. ⚡🔥⚡

Last update: Jan 09, 2023

Overview

Lightning-Hydra-Template

A clean and scalable template to kickstart your deep learning project 🚀 ⚡ 🔥
Click on Use this template to initialize new repository.

Suggestions are always welcome!

📌 Introduction

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: this template is not fitted to be a production environment, should be used more as a fast experimentation tool. Also, even though Lightning is very flexible, it's not well suited for every possible deep learning task. See #Limitations for more.

Why PyTorch Lightning?

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.

Why Hydra?

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.

Main Ideas Of This Template

Predefined Structure: clean and scalable so that work can easily be extended and replicated | #Project Structure
Rapid Experimentation: thanks to automating pipeline with config files and hydra command line superpowers | #Your Superpowers
Reproducibility: obtaining similar results is supported in multiple ways | #Reproducibility
Little Boilerplate: so pipeline can be easily modified | #How It Works
Main Configuration: main config file specifies default training configuration | #Main Project Configuration
Experiment Configurations: can be composed out of smaller configs and override chosen hyperparameters | #Experiment Configuration
Workflow: comes down to 4 simple steps | #Workflow
Experiment Tracking: many logging frameworks can be easily integrated, like Tensorboard, MLFlow or W&B | #Experiment Tracking
Logs: all logs (checkpoints, data from loggers, hparams, etc.) are stored in a convenient folder structure imposed by Hydra | #Logs
Hyperparameter Search: made easier with Hydra built-in plugins like Optuna Sweeper | #Hyperparameter Search
Tests: unit tests and shell/command based tests for speeding up the development | #Tests
Best Practices: a couple of recommended tools, practices and standards for efficient workflow and reproducibility | #Best Practices

Project Structure

The directory structure of new project looks like this:

├── bash                    <- Bash scripts
│   └── 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
│   ├── local                   <- Local configs
│   ├── logger                  <- Logger configs
│   ├── mode                    <- Running mode 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
│   ├── vendor                  <- Third party code that cannot be installed using PIP/Conda
│   │
│   └── 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

🚀 Quickstart

# clone project
git clone https://github.com/ashleve/lightning-hydra-template
cd lightning-hydra-template

# [OPTIONAL] create conda environment
conda create -n myenv python=3.8
conda activate myenv

# install pytorch according to instructions
# https://pytorch.org/get-started/

# install requirements
pip install -r requirements.txt

Template contains example with MNIST classification.
When running python run.py you should see something like this:

⚡ Your Superpowers

(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-4

You 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.strategy=ddp

# train with DDP (Distributed Data Parallel) (8 GPUs, 2 nodes)
python run.py trainer.gpus=4 +trainer.num_nodes=2 +trainer.strategy=ddp

Train with mixed precision

# train with pytorch native automatic mixed precision (AMP)
python run.py trainer.gpus=1 +trainer.precision=16

Train model with any logger available in PyTorch Lightning, like Weights&Biases or Tensorboard

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=wandb

Use 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 loggers
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_simple

Attach 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=default

Use 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 in debug mode
# changes logging folder to `logs/debug/...`
# enables trainer debugging options specified in `configs/trainer/debug.yaml`
# sets level of all command line loggers to 'DEBUG'
python run.py mode=debug

# enable trainer debugging options specified in `configs/trainer/debug.yaml`
python run.py trainer=debug

# run 1 train, val and test loop, using only 1 batch
python run.py +trainer.fast_dev_run=true

# raise exception if there are any numerical anomalies in tensors, like NaN or +/-inf
python run.py +trainer.detect_anomaly=true

# 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=2

Resume 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

⚠️ This sweep is not failure resistant (if one job crashes than the whole sweep crashes).

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_simple

⚠️ Currently this sweep is not failure resistant (if one job crashes than the whole sweep crashes). Might be supported in future Hydra release.

Execute 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 tab key. Learn more here.

🐳 Docker

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>

❤️ Contributions

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!

ℹ️ Guide

How To Get Started

First, you should probably get familiar with PyTorch Lightning
Next, go through Hydra quick start guide and basic Hydra tutorial

How It Works

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: 0.001

Using 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.

Main Project Configuration

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
# 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: True

Experiment Configuration

Location: 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 /mode: exp.yaml
  - 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

# name of the run determines folder name in logs and is accessed by loggers
name: "example_simple"

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: 64

Show advanced example

# to execute this experiment run:
# python run.py experiment=example_full

defaults:
  - override /mode: exp.yaml
  - 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

name: "example_full"

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"
    name: ${name}
    tags: ["best_model", "mnist"]
    notes: "Description of this model."

Local Configuration

Location: configs/local
Some configurations are user/machine/installation specific (e.g. configuration of local cluster, or harddrive paths on a specific machine). For such scenarios, a file configs/local/default.yaml can be created which is automatically loaded but not tracked by Git.

Local Slurm cluster config example

# @package _global_

defaults:
  - override /hydra/launcher@_here_: submitit_slurm

data_dir: /mnt/scratch/data/

hydra:
  launcher:
    timeout_min: 1440
    gpus_per_task: 1
    gres: gpu:1
  job:
    env_set:
      MY_VAR: /home/user/my/system/path
      MY_KEY: asdgjhawi8y23ihsghsueity23ihwd

Workflow

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

Logs

Hydra creates new working directory for every executed run.
This means your working directory is different for every run, which might not be compatible with some libraries and workflows. By default, logs have the following structure:

├── logs
│   ├── experiments           # Folder for logs generated by experiments
│   │   ├── experiment_name     # Name of the experiment
│   │   │   ├── runs
│   │   │   │   ├── YYYY-MM-DD        	# Date of execution
│   │   │   │   │   ├── HH-MM-SS          # Time of execution
│   │   │   │   │   │   ├── .hydra          # Hydra logs
│   │   │   │   │   │   ├── wandb           # Weights&Biases logs
│   │   │   │   │   │   ├── checkpoints     # Training checkpoints
│   │   │   │   │   │   └── ...             # Any other thing saved during training
│   │   │   │   │   └── ...
│   │   │   │   └── ...
│   │   │   │
│   │   │   └── multiruns
│   │   │       ├── YYYY-MM-DD
│   │   │       │   ├── HH-MM-SS
│   │   │       │   │   ├── 1               # Multirun job number
│   │   │       │   │   │   └── ...
│   │   │       │   │   ├── 2
│   │   │       │   │   └── ...
│   │   │       │   └── ...
│   │   │       └── ...
│   │   └── ...
│   │
│   ├── debugs                  # Folder for logs generated during debugging
│   │ 	├── runs
│   │   │   ├── YYYY-MM-DD
│   │   │   │   ├── HH-MM-SS
│   │   │   │   └── ...
│   │   │   └── ...
│   │   └── multiruns
│   │       ├── YYYY-MM-DD
│   │       │   ├── HH-MM-SS
│   │       │   └── ...
│   │       └─ ...
│   │
│   ├── runs                    # Folder for logs generated by normal runs
│   │   ├── YYYY-MM-DD
│   │   │   ├── HH-MM-SS
│   │   │   └── ...
│   │   └── ...
│   │
│   └── multiruns               # Folder for logs generated by normal multiruns (sweeps)
│       ├── YYYY-MM-DD
│       |   ├── HH-MM-SS
│       |   └── ...
│       └── ...
│

You can change this structure by modifying paths in hydra configuration.

Experiment Tracking

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_name

You 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.

Hyperparameter Search

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_best"

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.

Inference

The following code is an example of loading model from checkpoint and running predictions.

Show 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()

Tests

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"

To speed up the development, you can once in a while execute tests that run a couple of quick experiments, like training 1 epoch on 25% of data, executing single train/val/test step, etc. Those kind of tests don't check for any specific output - they exist to simply verify that executing some bash 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 1 epoch is too much for your model, then make it run for a couple of batches instead (by using the right trainer flags).

Callbacks

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=wandb

To see the result of all the callbacks attached, take a look at this experiment dashboard.

Multi-GPU Training

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.strategy=ddp

⚠️ When using DDP you have to be careful how you write your models - learn more here.

Reproducibility

What provides reproducibility:

Hydra manages your configs
Hydra manages your logging paths and makes every executed run store its hyperparameters and config overrides in a separate file in logs
Single seed for random number generators in pytorch, numpy and python.random
LightningDataModule allows you to encapsulate data split, transformations and default parameters in a single, clean abstraction
LightningModule separates your research code from engineering code in a clean way
Experiment tracking frameworks take care of logging metrics and hparams, some can also store results and artifacts in cloud
Pytorch Lightning takes care of creating training checkpoints
Example callbacks for wandb show how you can save and upload a snapshot of codebase every time the run is executed, as well as upload ckpts and track model gradients

You can load the config of previous run using:

python run.py --config-path /logs/runs/.../.hydra/ --config-name config.yaml

The config.yaml from .hydra folder contains all overriden parameters and sections. This approach however is not officially supported by Hydra and doesn't override the hydra/ part of the config, meaning logging paths will revert to default!

Limitations

Currently, template doesn't support k-fold cross validation, but it's possible to achieve it with Lightning Loop interface. See the official example. Implementing it requires rewriting the training pipeline.
Pytorch Lightning might not be the best choice for scalable reinforcement learning, it's probably better to use something like Ray.
Currently hyperparameter search with Hydra Optuna Plugin doesn't support prunning.
Hydra changes working directory to new logging folder for every executed run, which might not be compatible with the way some libraries work.

Useful Tricks

Accessing datamodule attributes in model

The simplest way is to pass datamodule attribute directly to model on initialization:
```
# ./src/train.py
datamodule = hydra.utils.instantiate(config.datamodule)
model = hydra.utils.instantiate(config.model, some_param=datamodule.some_param)
```
This is not a very robust solution, since it assumes all your datamodules have some_param attribute available (otherwise the run will crash).

If you only want to access datamodule config, you can simply pass it as an init parameter:

# ./src/train.py
model = hydra.utils.instantiate(config.model, dm_conf=config.datamodule, _recursive_=False)

Now you can access any datamodule config part like this:

# ./src/models/my_model.py
class MyLitModel(LightningModule):
	def __init__(self, dm_conf, param1, param2):
		super().__init__()

		batch_size = dm_conf.batch_size

If you need to access the datamodule object attributes, a little hacky solution is to add Omegaconf resolver to your datamodule:

# ./src/datamodules/my_datamodule.py
from omegaconf import OmegaConf

class MyDataModule(LightningDataModule):
	def __init__(self, param1, param2):
		super().__init__()

		self.param1 = param1

		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 return attribute 'param1' from datamodule object
param1: ${datamodule: param1}

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 or it will crash. You also need to set resolve=False in print_config() in run.py or it will throw errors:

# ./src/run.py
utils.print_config(config, resolve=False)

Automatic 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. Add some commands to your .autoenv file, like in the example below:

# 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)

To setup this automation for bash, execute the following line (it will append your .bashrc file):

> ~/.bashrc">

echo "autoenv() { if [ -x .autoenv ]; then source .autoenv ; echo '.autoenv executed' ; fi } ; cd() { builtin cd \"\$@\" ; autoenv ; } ; autoenv" >> ~/.bashrc

Lastly add execution previliges to your .autoenv file:
```
chmod +x .autoenv
```
(for safety, only .autoenv with previligies will be executed)

Explanation

The mentioned line appends your .bashrc file with 2 commands:

autoenv() { if [ -x .autoenv ]; then source .autoenv ; echo '.autoenv executed' ; fi } - this declares the autoenv() function, which executes .autoenv file if it exists in current work dir and has execution previligies
cd() { builtin cd \"\$@\" ; autoenv ; } ; autoenv - this extends behaviour of cd command, to make it execute autoenv() function each time you change folder in terminal or open new terminal

Best Practices

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. It also allows you to acccess your environments globally.

Example installation:

wget https://repo.anaconda.com/miniconda/Miniconda3-latest-Linux-x86_64.sh
bash Miniconda3-latest-Linux-x86_64.sh

Create new conda environment:

conda create -n myenv python=3.8
conda activate myenv

Use 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-commit

Next, install hooks from .pre-commit-config.yaml:

pre-commit install

After 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 -a

Set 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_path

All 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 init

To start tracking a file or directory, use dvc add:

dvc add data/MNIST

DVC 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:

=1.10.0", "pytorch-lightning>=1.4.0", "hydra-core>=1.1.0", ], packages=find_packages(), )">

from setuptools import find_packages, setup


setup(
    name="src",  # change "src" folder name to your project name
    version="0.0.0",
    description="Describe Your Cool Project",
    author="...",
    author_email="...",
    url="https://github.com/ashleve/lightning-hydra-template",  # replace with your own github project link
    install_requires=[
        "pytorch>=1.10.0",
        "pytorch-lightning>=1.4.0",
        "hydra-core>=1.1.0",
    ],
    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 --upgrade

So 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 MNISTDataModule

Other Repositories

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

License

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

Your Project Name

Description

What it does

How to run

Install dependencies

# clone project
git clone https://github.com/YourGithubName/your-repo-name
cd your-repo-name

# [OPTIONAL] create conda environment
conda create -n myenv python=3.8
conda activate myenv

# install pytorch according to instructions
# https://pytorch.org/get-started/

# install requirements
pip install -r requirements.txt

Train model with default configuration

# train on CPU
python run.py trainer.gpus=0

# train on GPU
python run.py trainer.gpus=1

Train model with chosen experiment configuration from configs/experiment/

python run.py experiment=experiment_name.yaml

You can override any parameter from command line like this

python run.py trainer.max_epochs=20 datamodule.batch_size=64

Comments

MultiGPU Error

(update to indicate the bug version is V1.1)

Thanks for your awesome work!

Reproduce

Directly download the main code (ReleaseV1.1) and execute this line(or any line using multiple GPUs)

python run.py trainer.gpus=[0,1]

Environment

Machine

1080Ti x 4, Ubuntu18.04

conda env showed with `pip list`

Package                           Version
--------------------------------- ---------
absl-py                           0.15.0
aiohttp                           3.8.0
aiosignal                         1.2.0
alembic                           1.6.5
antlr4-python3-runtime            4.8
anyio                             3.3.4
argon2-cffi                       21.1.0
async-timeout                     4.0.0
attrs                             21.2.0
autopage                          0.4.0
Babel                             2.9.1
backcall                          0.2.0
backports.entry-points-selectable 1.1.0
backports.functools-lru-cache     1.6.4
black                             21.10b0
bleach                            4.1.0
brotlipy                          0.7.0
cachetools                        4.2.4
certifi                           2021.10.8
cffi                              1.15.0
cfgv                              3.3.1
chardet                           4.0.0
charset-normalizer                2.0.7
click                             8.0.3
cliff                             3.9.0
cmaes                             0.8.2
cmd2                              2.2.0
colorama                          0.4.4
colorlog                          6.5.0
commonmark                        0.9.1
conda                             4.10.3
conda-package-handling            1.7.3
configparser                      5.0.2
cryptography                      35.0.0
cycler                            0.10.0
debugpy                           1.5.1
decorator                         5.1.0
defusedxml                        0.7.1
distlib                           0.3.3
docker-pycreds                    0.4.0
entrypoints                       0.3
filelock                          3.3.1
flake8                            4.0.1
frozenlist                        1.2.0
fsspec                            2021.10.1
future                            0.18.2
gitdb                             4.0.9
GitPython                         3.1.24
google-auth                       2.3.3
google-auth-oauthlib              0.4.6
googledrivedownloader             0.4
greenlet                          1.1.2
grpcio                            1.41.0
hydra-colorlog                    1.1.0
hydra-core                        1.1.1
hydra-optuna-sweeper              1.1.1
identify                          2.3.3
idna                              3.3
importlib-resources               5.4.0
iniconfig                         1.1.1
ipykernel                         6.4.1
ipython                           7.28.0
ipython-genutils                  0.2.0
ipywidgets                        7.6.5
isodate                           0.6.0
isort                             5.9.3
jedi                              0.18.0
Jinja2                            3.0.2
joblib                            1.1.0
json5                             0.9.6
jsonschema                        4.1.0
jupyter-client                    7.0.6
jupyter-core                      4.8.1
jupyter-server                    1.11.1
jupyterlab                        3.2.0
jupyterlab-pygments               0.1.2
jupyterlab-server                 2.8.2
jupyterlab-widgets                1.0.2
kiwisolver                        1.3.2
llvmlite                          0.37.0
Mako                              1.1.5
mamba                             0.17.0
Markdown                          3.3.4
MarkupSafe                        2.0.1
matplotlib                        3.4.3
matplotlib-inline                 0.1.3
mccabe                            0.6.1
mistune                           0.8.4
mkl-fft                           1.3.0
mkl-random                        1.2.2
mkl-service                       2.4.0
msgpack                           1.0.2
multidict                         5.2.0
mypy-extensions                   0.4.3
nbclassic                         0.3.2
nbclient                          0.5.4
nbconvert                         6.2.0
nbformat                          5.1.3
nest-asyncio                      1.5.1
networkx                          2.6.3
nodeenv                           1.6.0
notebook                          6.4.4
numba                             0.54.1
numpy                             1.20.3
oauthlib                          3.1.1
olefile                           0.46
omegaconf                         2.1.1
optuna                            2.10.0
packaging                         21.0
pandas                            1.3.4
pandocfilters                     1.5.0
parso                             0.8.2
pathspec                          0.9.0
pathtools                         0.1.2
pbr                               5.6.0
pexpect                           4.8.0
pickleshare                       0.7.5
Pillow                            8.3.1
pip                               21.3.1
platformdirs                      2.4.0
plotly                            5.3.1
pluggy                            1.0.0
pre-commit                        2.15.0
prettytable                       2.2.1
prometheus-client                 0.11.0
promise                           2.3
prompt-toolkit                    3.0.20
protobuf                          3.18.1
psutil                            5.8.0
ptyprocess                        0.7.0
pudb                              2021.2.2
py                                1.10.0
pyasn1                            0.4.8
pyasn1-modules                    0.2.8
pycodestyle                       2.8.0
pycosat                           0.6.3
pycparser                         2.20
pyDeprecate                       0.3.1
pyflakes                          2.4.0
Pygments                          2.10.0
PyGSP                             0.5.1
pyOpenSSL                         21.0.0
pyparsing                         2.4.7
pyperclip                         1.8.2
pyrsistent                        0.18.0
PySocks                           1.7.1
pytest                            6.2.5
python-dateutil                   2.8.2
python-dotenv                     0.19.1
python-editor                     1.0.4
pytorch-lightning                 1.5.0
pytz                              2021.3
PyYAML                            6.0
pyzmq                             22.3.0
ray                               1.7.0
rdflib                            6.0.2
redis                             3.5.3
regex                             2021.11.1
requests                          2.26.0
requests-oauthlib                 1.3.0
requests-unixsocket               0.2.0
rich                              10.12.0
rsa                               4.7.2
ruamel-yaml-conda                 0.15.80
scikit-learn                      1.0
scikit-sparse                     0.4.6
scikit-umfpack                    0.3.2
scipy                             1.7.1
seaborn                           0.11.2
Send2Trash                        1.8.0
sentry-sdk                        1.4.3
setuptools                        58.0.4
sh                                1.14.2
shortuuid                         1.0.1
six                               1.16.0
sklearn                           0.0
smmap                             5.0.0
sniffio                           1.2.0
SQLAlchemy                        1.4.26
stevedore                         3.5.0
subprocess32                      3.5.4
tenacity                          8.0.1
tensorboard                       2.7.0
tensorboard-data-server           0.6.1
tensorboard-plugin-wit            1.8.0
termcolor                         1.1.0
terminado                         0.12.1
testpath                          0.5.0
thgsp                             0.1.0
threadpoolctl                     3.0.0
toml                              0.10.2
tomli                             1.2.2
torch                             1.8.2
torch-cluster                     1.5.9
torch-geometric                   2.0.1
torch-scatter                     2.0.8
torch-sparse                      0.6.12
torch-spline-conv                 1.2.1
torchaudio                        0.8.2
torchmetrics                      0.5.1
torchvision                       0.9.2
tornado                           6.1
tqdm                              4.62.3
traitlets                         5.1.0
typing-extensions                 3.10.0.2
urllib3                           1.26.7
urwid                             2.1.2
urwid-readline                    0.13
virtualenv                        20.10.0
wandb                             0.12.6
wcwidth                           0.2.5
webencodings                      0.5.1
websocket-client                  1.2.1
Werkzeug                          2.0.2
wheel                             0.37.0
widgetsnbextension                3.5.1
yacs                              0.1.8
yarl                              1.7.2
yaspin                            2.1.0
zipp                              3.6.0

Error Info

raceback (most recent call last):
  File "<string>", line 1, in <module>
  File "/home/amax/anaconda3/envs/pyg18/lib/python3.8/multiprocessing/spawn.py", line 116, in spawn_main
    exitcode = _main(fd, parent_sentinel)
  File "/home/amax/anaconda3/envs/pyg18/lib/python3.8/multiprocessing/spawn.py", line 125, in _main
    prepare(preparation_data)
  File "/home/amax/anaconda3/envs/pyg18/lib/python3.8/multiprocessing/spawn.py", line 236, in prepare
    _fixup_main_from_path(data['init_main_from_path'])
  File "/home/amax/anaconda3/envs/pyg18/lib/python3.8/multiprocessing/spawn.py", line 287, in _fixup_main_from_path
    main_content = runpy.run_path(main_path,
  File "/home/amax/anaconda3/envs/pyg18/lib/python3.8/runpy.py", line 264, in run_path
    code, fname = _get_code_from_file(run_name, path_name)
  File "/home/amax/anaconda3/envs/pyg18/lib/python3.8/runpy.py", line 234, in _get_code_from_file
    with io.open_code(decoded_path) as f:
FileNotFoundError: [Errno 2] No such file or directory: '/data/dbw/projects/lightning-hydra-template-main/logs/runs/2021-11-04/10-14-23/run.py'

bug

opened by bwdeng20 11

Code quality workflow

Hi, Thank you for the great template, I use it a lot 😁

I enhanced the check code formatting to run only on the changed files, and move it to separate a workflow file. From experience, sometimes pre-commit's output is large, so IMHO the separation is needed :)

I also fixed a small English mistake.

Thank you, Eli

Edit: consider badging the repository with pre-commit ;)

opened by elisim 9
`mode=debug` is not overriding the defaults list when running alongside an experiment config

When I use python run.py experiment=something mode=debug, I expect the trainer to be overridden by the debug.yaml one, but in fact this doesn't happen (looking at the config tree reveals that the default.yaml is being used. Additionally, I am also trying to disable loggers and callbacks in this mode, by setting them to null or none.yaml. However, they are also still being loaded.

I assume that this is because the experiment config also has an - override /trainer, but my knowledge of hydra is not enough to figure out why this isn't working.
enhancement investigation

opened by mariomeissner 9
hydra + multirun + ddp error

Hello. I found that the combination of multi-run and DDP doesn't work with the following command.

python run.py -m datamodule.batch_size=32,64 trainer=ddp trainer.max_epochs=2

Can you have a look at this?
bug

opened by shim94kr 9
Instantiate model using Hydra
Hydra allows us to recursively instantiate classes, which means we don't need to pass SimpleDenseNet parameters to MNISTLitModel, but can let Hydra create the class for us.

This cleans up the code, makes it more modular and reusable, and neatly groups the parameters together.

This leaves one open question though: what happens in

self.save_hyperparameters(logger=False)

Does it pick up on those nested hyper parameters?
opened by nils-werner 9
Multi-GPU training on DDP mode will log multiple times

I notice when I use +trainer.accelerator="ddp", the log file have repeated logs(The number of repeats is equal to the number of GPUs), refer to official docs, "Lightning implementation of DDP calls your script under the hood multiple times with the correct environment variables:". So we need a solution to handle this problem since DDP mode is much faster than DDP_spawn.
bug enhancement help wanted

opened by LuoXin-s 9
Update wandb_callbacks.py

UploadCodeAsArtifact: Add the functionality to use git to decide which files are source files. Can upload all files that are not ignored by git instead of all '*.py' file. UploadCodeAsArtifact: Add @rank_zero_only UploadCheckpointsAsArtifact: Add @rank_zero_only UploadCheckpointsAsArtifact: ckpts are output of the run, so use experiment.log_artifact(ckpts) instead

opened by zhengyu-yang 8
Multi-GPU bugs, AttributeError: Can't pickle local object 'log_hyperparameters..'

When I use Multi-GPU with 4 3090, I ran into AttributeError: Can't pickle local object 'log_hyperparameters..', it seems due to trainer.logger.log_hyperparams = lambda params: None trick in log_hyperparameters.

Traceback (most recent call last): File "/ghome/luoxin/projects/liif-lightning-hydra/run.py", line 34, in main return train(config) File "/ghome/luoxin/projects/liif-lightning-hydra/src/train.py", line 78, in train trainer.fit(model=model, datamodule=datamodule) File "/opt/conda/lib/python3.8/site-packages/pytorch_lightning/trainer/trainer.py", line 499, in fit self.dispatch() File "/opt/conda/lib/python3.8/site-packages/pytorch_lightning/trainer/trainer.py", line 546, in dispatch self.accelerator.start_training(self) File "/opt/conda/lib/python3.8/site-packages/pytorch_lightning/accelerators/accelerator.py", line 73, in start_training self.training_type_plugin.start_training(trainer) File "/opt/conda/lib/python3.8/site-packages/pytorch_lightning/plugins/training_type/ddp_spawn.py", line 108, in start_training mp.spawn(self.new_process, **self.mp_spawn_kwargs) File "/opt/conda/lib/python3.8/site-packages/torch/multiprocessing/spawn.py", line 230, in spawn return start_processes(fn, args, nprocs, join, daemon, start_method='spawn') File "/opt/conda/lib/python3.8/site-packages/torch/multiprocessing/spawn.py", line 179, in start_processes process.start() File "/opt/conda/lib/python3.8/multiprocessing/process.py", line 121, in start self._popen = self._Popen(self) File "/opt/conda/lib/python3.8/multiprocessing/context.py", line 284, in _Popen return Popen(process_obj) File "/opt/conda/lib/python3.8/multiprocessing/popen_spawn_posix.py", line 32, in init super().init(process_obj) File "/opt/conda/lib/python3.8/multiprocessing/popen_fork.py", line 19, in init self._launch(process_obj) File "/opt/conda/lib/python3.8/multiprocessing/popen_spawn_posix.py", line 47, in _launch reduction.dump(process_obj, fp) File "/opt/conda/lib/python3.8/multiprocessing/reduction.py", line 60, in dump ForkingPickler(file, protocol).dump(obj) AttributeError: Can't pickle local object 'log_hyperparameters..'
bug

opened by LuoXin-s 7
Testing best model checkpoint after training model with multiple GPUs

Hello.

First off, I am a huge fan of this project and want to thank its maintainers for their diligent efforts!

I noticed that in the default train.py example, the maintainers of this project would have users call trainer.test() on the best model checkpoint resulting from (single/multi) GPU training beforehand (if the test CLI argument is set to True). In such a case, I am a bit concerned about how Lightning will attempt to report its test results to users through the use of a logging tool like WandB. For example, I believe it is typically recommended to run a trained PyTorch/Lightning model on the test dataset(s) using only a single GPU to avoid any misreporting of test results.

Is such an option available with this project layout?
question

opened by amorehead 6
Can you share a method for NLP users to deal with Vocab Size?

Hi,

This is a great template!

However, I have yet to find a way to deal with vocab_size which is derived from datasets and cannot be preset in advance.

Do you know an elegant way to do it?
question

opened by yipliu 6
Refactor dockerfile

Hello, I am learning PyTorch Lightning and Hydra and come across finding this repo. Really useful and great work!

I made some changes in Dockerfile so the changes on the host machine can reflect within the container. When developing/training a model, mounting the files into the container instead of copying is more convenient and time-saving as all the changes I made and all the generated logs will be saved in the host machine.

Do you think it worth merging to the main branch or should be added to the dockerfiles branch?

opened by HotThoughts 6
Pytorch 2.0

New feature Pytorch 2.0 compile Change pytorch-lightning(deprecated) to lightning Pytorch 1.13.1 compatible with python 3.11 Only blocked tests torchvision but they are working to compatible with Python 3.11 Deleted Python 3.7 tests(very deprecated)

opened by johnnynunez 2
Assign stage flag from `train.yaml`
There is a def setup(self, stage: Optional[str] = None): stage flag in the datamodule, will it be possible to initiate stage -> train or stage-> test from train.yaml If look in train.yaml

# set False to skip model training train: True # evaluate on test set, using best model weights achieved during training # lightning chooses best weights based on the metric specified in checkpoint callback test: True

Just to save memory
opened by abhijeetdhakane 0
`train.log` doesn't show training progress bar

Awesome project honestly! I'm using your project right now but having some issues while training with SLURM, since I can't see the terminal of the scheduled nodes, I can only check train.log or slurm.log; but both of those log doesn't show the training progress bar (stop at Start training or some DDP warnings); but there are still logs on tensorboard and new ckpts. I've tried to disable colorlogs too but it still doesn't show up.

opened by bomcon123456 0
Support Intel Arc

You can add new device: intel arc now is supported by Pytorch: https://github.com/intel/intel-extension-for-pytorch/releases/tag/v1.10.200%2Bgpu python -m pip install intel_extension_for_pytorch intel-extension-for-pytorch
enhancement

opened by johnnynunez 2
How to define a conditional search space ?
Hi!

I want to search the best optimizer for the given "mnist_example" from SGD and Adam. However, for SGD, I also want to know which momentum value is the best (which Adam doesn't need), but for Adam, I need search beta.

Therefore, how can we define such a hparams_search/name.yaml file?

hydra: ... sweeper: ... params: model.optimizer._target_: choice("torch.optim.SGD", "torch.optim.Adam") ......

Thanks so much!
enhancement question
opened by tianshuocong 8

Add Custom Dataloader Example into template

@ashleve this is an excellent template. But if you add some files, then it will be usable for a lot many audiences. I felt a little tedious for data modules. The template has an MNIST data module which is 'standard.' In many used cases people use custom Dataset, If you add those, then it will be more helpful.

Just an example I tried:

datamodule -> component -> dataloader.py

class FaceLandmarksDataset(Dataset):
    """Face Landmarks dataset."""

    def __init__(self, csv_file, root_dir, transform=None):
        """
        Args:
            csv_file (string): Path to the csv file with annotations.
            root_dir (string): Directory with all the images.
            transform (callable, optional): Optional transform to be applied
                on a sample.
        """
        self.landmarks_frame = pd.read_csv(csv_file)
        self.root_dir = root_dir
        self.transform = transform

    def __len__(self):
        return len(self.landmarks_frame)

    def __getitem__(self, idx):
        if torch.is_tensor(idx):
            idx = idx.tolist()

        img_name = os.path.join(self.root_dir,
                                self.landmarks_frame.iloc[idx, 0])
        image = io.imread(img_name)
        landmarks = self.landmarks_frame.iloc[idx, 1:]
        landmarks = np.array([landmarks])
        landmarks = landmarks.astype('float').reshape(-1, 2)
        sample = {'image': image, 'landmarks': landmarks}

        if self.transform:
            sample = self.transform(sample)

        return sample

from https://pytorch.org/tutorials/beginner/data_loading_tutorial.html

then datamodules -> XYZ_datamodule.py

A DataModule implements 5 key methods:
        def prepare_data(self):
            # things to do on 1 GPU/TPU (not on every GPU/TPU in DDP)
            # download data, pre-process, split, save to disk, etc...
        def setup(self, stage):
            # things to do on every process in DDP
            # load data, set variables, etc...
        def train_dataloader(self):
            # return train dataloader
        def val_dataloader(self):
            # return validation dataloader
        def test_dataloader(self):
            # return test dataloader
        def teardown(self):
            # called on every process in DDP
            # clean up after fit or test

I ran into errors while doing the above, my datamodule_XYZ.yaml like:

_target_:src.datamodules.XYZ_datamodule.XZYDataModule
datasets:
       train:
       _target_: src.datamodules.components.dataloader.FaceLandmarksDataset
       csv_file: ...
       root_dir: ...
       transform:
           _target_: ...

      val:

      test:

batches:
   train:10
   val:10
   test:10

...

enhancement

opened by abhijeetdhakane 3

Releases(v1.4.0)

v1.4.0(Jul 16, 2022)
What's Changed

Adapt template to hydra 1.2 - no more changing the working directory by default

Rename test.py and test.yaml to eval.py and eval.yaml (so as to avoid confusion with project tests)

Move train.py and eval.py inside src/

Add pyrootutils package for standardizing the project root setup in train.py and eval.py

Rename pipelines to tasks

Create a separate folder for tasks

Add task_name to main config, which determines hydra output folder path

Introduce @task_wrapper decorator for applying utilities before and after the task is executed

Standardize what is returned from tasks: Tuple[metric_dict, object_dict]

Add SimpleDenseNet config to model config with recursive instantiation

Add optimizer config to model config using _partial_: true

Remove _convert_=partial from trainer instantiation (no longer needed since recent lightning release)

Add ckpt_path to main config, trainer.fit() and trainer.test(), for compatibility with recent lightning release

Add resetting val_acc_best metric at the start of the training to prevent storing results from validation sanity checks

Add verifying logger is not None before logging hparams

Add tags to main config

Add prompting user to input tags when none are provide to utils.extras()

Remove experiment name (since tags and task_name are enough)

Rename config to cfg since it's the standard naming convention in hydra

Split utils into multiple files: utils.py, rich_utils.py, pylogger.py

Add utils.instantiate_callbacks() and utils.instantiate_loggers() to reduce the boilerplate in tasks

Add utils.get_metric_value() for safely retrieving optimized metric.

Rename utils.finish() to utils.close_loggers()

Move extra config utils to configs/extras/default.yaml

Replace deprecated trainer.gpus argument with trainer.accelerator and trainer.devices

Add separate trainer configs for GPU, CPU, simulating DDP on CPU and MPS accelerator (Accelerated PyTorch Training on Mac)

Add hydra.mode=MULTIRUN to mnist_optuna.yaml config (so using -m is no longer needed when attaching this config)

Update mnist_opuna.yaml for compatibility with new search space syntax in hydra 1.2

Disable callbacks in debug configs by default (fixes debug/overfit.yaml)

Disable hydra command line debug logger in debug configs by default

Split callbacks config into multiple files

Replace setup.cfg with pyproject.toml since it's a more versatile standard (PEP 518)

Add pre-commit hooks: pyupgrade (automatically upgrading python syntax to newer version), bandit (security linter), codespell (spelling linter), mdformat (markdown formatting)

Redesign testing and add tests covering ddp, multirun, loggers, resuming training and evaluation

Implement CI workflows with GitHub Actions: executing pytest, test code coverage measuring, code quality testing for main branch and PRs

Add depandabot

Add pull request template

Add setup.py

Add Makefile

Update README.md

Contributors

@nils-werner @johnnynunez @elisim @yu-xiang-wang @yipliu @Gxinhu @binlee52
Source code(tar.gz)
Source code(zip)
v1.3(Feb 19, 2022)
The template has been significantly refactored.

List of changes:

Introduce multiple pipelines, to showcase example of how one can separate training from evaluation, run.py has been replaced by train.py and test.py

The mode group config has been removed since it was confusing, now every run is treated as an experiment, and debugging is moved to a separate config group

Introduce debug config group

Introduce log_dir config group

Move wandb callbacks to the branch wandb-callbacks to make template logger-agnostic

Refactor rich config printing, now all config groups are always printed instead of just the pre-selected ones, but you can still decide on the print order

Add nbstripout to pre-commit hooks, for automatic clearing of jupyter notebooks outputs before commit

Update packages in requirements.txt and pre-commit-config.yaml to newest versions

Remove some of the unimportant default optuna parameters in mnist_optuna.yaml and add more explanatory comments

Remove no longer needed utilities from utils.extras()

Add config flag for skipping training

Fix hydra package versions in requirements.txt for mac compatibility

Remove redundant parts in filenames: mnist_model.yaml -> mnist.yaml, mnist_datamodule.yaml -> datamodule.yaml

Change mnist_model.py -> mnist_module.py and MNISTLitModel -> MNISTLitModule

Rename folder modules/ to components/

Change accelerator="ddp" to strategy="ddp" since it was depracated by lightning

Remove trainer arguments depracated by lightning: weight summary and progress_bar_refresh_rate

Specify black profile for isort inside .pre-commit-config.yaml just in case someone deletes the setup.cfg

Specify testpath in setup.cfg so pytest knows all test files are placed only in tests/ folder

Allow for using relative checkpoint paths

Rename folder bash to scripts

Introduce vendor dir as a "best practice" for storing third party code

Introduce local config files in configs/local/, which can be used for storing machine/user specific configurations, e.g. configuration of slurm cluster

Unify logging directories structure

Add RichModelSummary to default callbacks

Fix missing parameter in "Accessing datamodule attributes" trick in README.md

General README.md improvements

Special thanks to: @nils-werner @charlesincharge @Steve-Tod for their PRs.
Source code(tar.gz)
Source code(zip)
v1.2(Nov 19, 2021)
List of changes:

Update template for compatibility with lightning v1.5 and pytorch v1.10

General documentation improvements

Move LICENSE to README.md

Add manual resetting of metrics at the end of every epoch, to make sure no one makes hard to spot calculation mistakes

Add experiment mode to all experiment configs

Improve logging paths for experiment mode

Add MaxMetric to model, for computation of best so far validation accuracy

Add RichProgressBar to default callbacks for the pretty formatted progress bar

Get rid of the trick for preventing auto hparam logging, since lightning now supports it with self.save_hyperparameters(logger=False)

Add self.save_hyperparameters() to datamodule since lightinng now supports it

Deprecate Apex support since native pytorch mixed-precision is better

Deprecate bash script for conda setup since installation commands change too often to maintain it

Change trainer.terminate_on_nan debug option to trainer.detect_anomaly for compatibility with lightning v1.5

Specify model and datamodule during trainer.test(), for compatibility with lightning v1.5

Remove configs/trainer/all_params.yaml

Make hyperparameter optimization compatible with lightning v1.5

Specify that EarlyStopping patience is counted in validation epochs and not in training epochs.

Add a new way for accessing datamodule attributes to the README.md

Make debug mode automatically set the level of all command-line loggers to DEBUG

Make debug mode automatically set the trainer config to debug.yaml

Add generator seed to prevent test data leaking to train data in datamodule.setup() when seed is not set up

Move Dockerfile to dockerfiles branch

Modifiy configs/trainer/debug.yaml to enable some debug options

Remove unused if config.get("debug"): in extras

Special thanks for PRs to: @CharlesGaydon, @eungbean, @gscriva
Source code(tar.gz)
Source code(zip)
v1.1(Sep 28, 2021)
introduce different running modes: default, debug, experiment

fix pytorch installation in setup_conda.sh

fix incorrect calculation of precision, recall and f1 score in wandb callback

add _self_ to config.yaml for compatibility with hydra1.1

fix setting seed in train.py so it's skipped when seed=null

add exception message when trying to use wandb callbacks with trainer.fast_dev_run=true

change axis=-1 to dim=-1 in LogImagePredictions callback

add 'Reproducibilty' section to README.md

UploadCodeAsArtifact: now uploads all files that are not ignored by git instead of all *.py files.

UploadCheckpointsAsArtifact: now uses experiment.log_artifact(ckpts) and uploads also on keyboard interrupt

Source code(tar.gz)
Source code(zip)
v1.0(Jul 21, 2021)
update to Hydra 1.1

add bash folder with scripts for conda setup and run scheduling

add saving seed in log_hyperparameters() method

redesign Dockerfile to make it weight less

add test_after_training parameter to config

add inheritance to trainer configs

remove forcing ddp-friendly configuration

refactor tests

remove conda_env_gpu.yaml

remove default langage version from pre-commit config

add mnist datamodule unit test

rename test folder from 'smoke' to 'shell'

remove wandb import from utils.py

change 'use_artifact' to 'log_artifact' in wandb callbacks

add rank zero decorator to wandb callbacks

add dumping rich tree config to file

get rid of = character in ckpt names

update requirements.txt

update README.md

Source code(tar.gz)
Source code(zip)
v0.9(May 21, 2021)

Source code(tar.gz)
Source code(zip)

PyTorch Lightning + Hydra. A feature-rich template for rapid, scalable and reproducible ML experimentation with best practices. ⚡🔥⚡

Related tags

Overview

Lightning-Hydra-Template

📌 Introduction

Why PyTorch Lightning?

Why Hydra?

Main Ideas Of This Template

Project Structure

🚀 Quickstart

⚡ Your Superpowers

🐳 Docker

❤️ Contributions

ℹ️ Guide

How To Get Started

How It Works

Main Project Configuration

Experiment Configuration

Local Configuration

Workflow

Logs

Experiment Tracking

Hyperparameter Search

Inference

Tests

Callbacks

Multi-GPU Training

Reproducibility

Limitations

Useful Tricks

Best Practices

Other Repositories

License

Your Project Name

Description

How to run

Comments

Thanks for your awesome work!

Reproduce

Environment

Machine

conda env showed with pip list

Error Info

Releases(v1.4.0)

v1.4.0(Jul 16, 2022)

What's Changed

Contributors

v1.3(Feb 19, 2022)

v1.2(Nov 19, 2021)

v1.1(Sep 28, 2021)

v1.0(Jul 21, 2021)

v0.9(May 21, 2021)

Owner

Łukasz Zalewski

Official implementation for “Unsupervised Low-Light Image Enhancement via Histogram Equalization Prior”

CNN designed for pansharpening

TeST: Temporal-Stable Thresholding for Semi-supervised Learning

PyTorch implementation of PSPNet segmentation network

An index of recommendation algorithms that are based on Graph Neural Networks.

The official repository for our paper "The Devil is in the Detail: Simple Tricks Improve Systematic Generalization of Transformers". We significantly improve the systematic generalization of transformer models on a variety of datasets using simple tricks and careful considerations.

Implementation of TransGanFormer, an all-attention GAN that combines the finding from the recent GanFormer and TransGan paper

Iranian Cars Detection using Yolov5s, PyTorch

A compendium of useful, interesting, inspirational usage of pandas functions, each example will be an ipynb file

Demystifying How Self-Supervised Features Improve Training from Noisy Labels

Beta Shapley: a Unified and Noise-reduced Data Valuation Framework for Machine Learning

Filtering variational quantum algorithms for combinatorial optimization

Official repository of the paper "A Variational Approximation for Analyzing the Dynamics of Panel Data". Mixed Effect Neural ODE. UAI 2021.

Unsupervised MRI Reconstruction via Zero-Shot Learned Adversarial Transformers

Autoencoders pretraining using clustering

Small repo describing how to use Hugging Face's Wav2Vec2 with PyCTCDecode

Team nan solution repository for FPT data-centric competition. Data augmentation, Albumentation, Mosaic, Visualization, KNN application

Deep Unsupervised 3D SfM Face Reconstruction Based on Massive Landmark Bundle Adjustment.

Official implementation of VQ-Diffusion

Source code of AAAI 2022 paper "Towards End-to-End Image Compression and Analysis with Transformers".

conda env showed with `pip list`