catalyst

Accelerated deep learning R&D

3,272

386

3,272

View on GitHub

Top Related Projects

ignite

4,509

High-level library to help with training and evaluating neural networks in PyTorch flexibly and transparently.

ray

32,953

Ray is a unified framework for scaling AI and Python applications. Ray consists of a core distributed runtime and a set of AI Libraries for accelerating ML workloads.

DeepSpeed

34,658

DeepSpeed is a deep learning optimization library that makes distributed training and inference easy, efficient, and effective.

horovod

14,147

Distributed training framework for TensorFlow, Keras, PyTorch, and Apache MXNet.

mlflow

18,287

Open source platform for the machine learning lifecycle

wandb

8,869

The AI developer platform. Use Weights & Biases to train and fine-tune models, and manage models from experimentation to production.

Quick Overview

Catalyst is a PyTorch framework for Deep Learning research and development. It focuses on reproducibility, rapid experimentation, and code/ideas reuse. Catalyst provides a high-level API for training loops, allowing researchers and developers to focus on the model architecture and problem-solving rather than boilerplate code.

Pros

Highly customizable and flexible, allowing for easy experimentation
Provides a rich set of callbacks and metrics for monitoring and optimizing training
Supports distributed training and mixed precision out of the box
Integrates well with other popular libraries like Optuna for hyperparameter tuning

Cons

Steeper learning curve compared to some other PyTorch wrappers
Documentation can be inconsistent or lacking in some areas
May have more overhead for simple projects compared to vanilla PyTorch
Smaller community compared to some other deep learning frameworks

Code Examples

Basic training loop with Catalyst:

from catalyst import dl

# Define your model, criterion, optimizer, and datasets
model = MyModel()
criterion = nn.CrossEntropyLoss()
optimizer = torch.optim.Adam(model.parameters())
train_dataset = MyDataset()
valid_dataset = MyDataset()

# Create Catalyst Runner
runner = dl.SupervisedRunner()

# Run the training
runner.train(
    model=model,
    criterion=criterion,
    optimizer=optimizer,
    loaders={
        "train": DataLoader(train_dataset),
        "valid": DataLoader(valid_dataset)
    },
    num_epochs=10,
    callbacks=[dl.AccuracyCallback()]
)

Custom callback for logging:

from catalyst import dl

class MyCallback(dl.Callback):
    def on_epoch_end(self, runner):
        logger.info(f"Epoch {runner.epoch} ended")

runner.train(
    # ... other parameters ...
    callbacks=[MyCallback()]
)

Distributed training with Catalyst:

from catalyst.dl import DDPRunner

runner = DDPRunner()

runner.train(
    # ... other parameters ...
    num_gpus=4  # Specify the number of GPUs to use
)

Getting Started

To get started with Catalyst, first install it using pip:

pip install catalyst

Then, import the necessary modules and create a basic training script:

from catalyst import dl
import torch
from torch import nn
from torch.utils.data import DataLoader, TensorDataset

# Create a simple model and dataset
model = nn.Linear(10, 1)
dataset = TensorDataset(torch.randn(100, 10), torch.randn(100, 1))
loader = DataLoader(dataset, batch_size=32)

# Define runner and train
runner = dl.SupervisedRunner()
runner.train(
    model=model,
    criterion=nn.MSELoss(),
    optimizer=torch.optim.Adam(model.parameters()),
    loaders={"train": loader},
    num_epochs=5,
    verbose=True
)

This basic example demonstrates how to set up a simple training loop with Catalyst. For more advanced usage, refer to the official documentation and examples.

Competitor Comparisons

ignite

4,509

High-level library to help with training and evaluating neural networks in PyTorch flexibly and transparently.

Pros of Ignite

More lightweight and flexible, allowing for easier customization
Better integration with PyTorch ecosystem and tools
More extensive documentation and examples

Cons of Ignite

Less built-in functionality for complex tasks
Steeper learning curve for beginners
Fewer pre-built metrics and callbacks

Code Comparison

Catalyst:

from catalyst import dl

runner = dl.SupervisedRunner()
runner.train(
    model=model,
    criterion=criterion,
    optimizer=optimizer,
    loaders=loaders,
    num_epochs=10,
    callbacks=[dl.AccuracyCallback()]
)

Ignite:

from ignite.engine import create_supervised_trainer, create_supervised_evaluator
from ignite.metrics import Accuracy

trainer = create_supervised_trainer(model, optimizer, criterion)
evaluator = create_supervised_evaluator(model, metrics={'accuracy': Accuracy()})

trainer.run(train_loader, max_epochs=10)

Both Catalyst and Ignite provide high-level APIs for training PyTorch models, but Ignite offers more granular control over the training process. Catalyst's API is more concise and easier to use for standard tasks, while Ignite's approach allows for greater customization at the cost of verbosity.

ray

32,953

Ray is a unified framework for scaling AI and Python applications. Ray consists of a core distributed runtime and a set of AI Libraries for accelerating ML workloads.

Pros of Ray

More comprehensive distributed computing framework, supporting various workloads beyond machine learning
Larger community and ecosystem, with extensive documentation and third-party integrations
Better scalability for large-scale distributed applications and cluster management

Cons of Ray

Steeper learning curve due to its broader scope and more complex API
Potentially overkill for simpler machine learning projects that don't require distributed computing
Higher resource overhead for small-scale tasks

Code Comparison

Ray example:

import ray

@ray.remote
def f(x):
    return x * x

futures = [f.remote(i) for i in range(4)]
print(ray.get(futures))

Catalyst example:

from catalyst import dl

class CustomRunner(dl.Runner):
    def predict_batch(self, batch):
        return self.model(batch)

runner = CustomRunner()

DeepSpeed

34,658

DeepSpeed is a deep learning optimization library that makes distributed training and inference easy, efficient, and effective.

Pros of DeepSpeed

Highly optimized for large-scale distributed training, especially for transformer models
Offers advanced features like ZeRO optimizer stages and pipeline parallelism
Seamless integration with popular frameworks like PyTorch and Hugging Face Transformers

Cons of DeepSpeed

Steeper learning curve due to its focus on advanced optimization techniques
May be overkill for smaller projects or simpler model architectures
Less emphasis on experiment tracking and logging compared to Catalyst

Code Comparison

DeepSpeed:

model_engine, optimizer, _, _ = deepspeed.initialize(
    args=args,
    model=model,
    model_parameters=params
)

Catalyst:

runner = dl.SupervisedRunner()
runner.train(
    model=model,
    criterion=criterion,
    optimizer=optimizer,
    loaders=loaders,
    num_epochs=10,
    callbacks=[dl.AccuracyCallback()]
)

DeepSpeed focuses on distributed training initialization, while Catalyst provides a higher-level API for training loops and experiment management. DeepSpeed is more suitable for large-scale projects requiring advanced optimization, whereas Catalyst offers a more user-friendly approach for general deep learning experiments.

horovod

14,147

Distributed training framework for TensorFlow, Keras, PyTorch, and Apache MXNet.

Pros of Horovod

Specializes in distributed deep learning, offering excellent scalability across multiple GPUs and nodes
Supports multiple deep learning frameworks (TensorFlow, PyTorch, MXNet)
Integrates well with popular cloud platforms and cluster managers

Cons of Horovod

Steeper learning curve, especially for those new to distributed training
Requires more setup and configuration compared to Catalyst
Less focus on experiment management and logging features

Code Comparison

Horovod (distributed training):

import horovod.torch as hvd

hvd.init()
torch.cuda.set_device(hvd.local_rank())
optimizer = hvd.DistributedOptimizer(optimizer, named_parameters=model.named_parameters())
hvd.broadcast_parameters(model.state_dict(), root_rank=0)

Catalyst (experiment setup):

from catalyst import dl

runner = dl.SupervisedRunner()
runner.train(
    model=model,
    criterion=criterion,
    optimizer=optimizer,
    loaders=loaders,
    num_epochs=10,
    callbacks=[dl.AccuracyCallback()]
)

mlflow

18,287

Open source platform for the machine learning lifecycle

Pros of MLflow

Broader scope: MLflow covers experiment tracking, model management, and deployment, while Catalyst focuses primarily on training loops
Larger community and ecosystem: More integrations, plugins, and third-party support
Language-agnostic: Supports multiple programming languages, not just Python

Cons of MLflow

Steeper learning curve: More complex setup and configuration due to its broader feature set
Heavier weight: Requires more dependencies and resources compared to Catalyst's lightweight approach
Less opinionated: Provides more flexibility but less structure for organizing deep learning experiments

Code Comparison

MLflow:

import mlflow

mlflow.start_run()
mlflow.log_param("learning_rate", 0.01)
mlflow.log_metric("accuracy", 0.95)
mlflow.end_run()

Catalyst:

from catalyst import dl

runner = dl.SupervisedRunner()
runner.train(
    model=model,
    criterion=criterion,
    optimizer=optimizer,
    loaders=loaders,
    num_epochs=10,
    callbacks=[dl.AccuracyCallback()]
)

wandb

8,869

The AI developer platform. Use Weights & Biases to train and fine-tune models, and manage models from experimentation to production.

Pros of wandb

More comprehensive experiment tracking and visualization tools
Broader language and framework support beyond PyTorch
Larger community and more extensive documentation

Cons of wandb

Requires internet connection for full functionality
Potential privacy concerns with cloud-based storage of experiment data
Steeper learning curve for beginners compared to Catalyst

Code Comparison

wandb:

import wandb

wandb.init(project="my-project")
wandb.config.hyperparameters = {
    "learning_rate": 0.01,
    "epochs": 100,
}
wandb.log({"loss": 0.5, "accuracy": 0.8})

Catalyst:

from catalyst import dl

runner = dl.SupervisedRunner()
runner.train(
    model=model,
    criterion=criterion,
    optimizer=optimizer,
    loaders=loaders,
    num_epochs=100,
    logdir="./logs",
)

Both wandb and Catalyst offer tools for machine learning experiment tracking and management. wandb provides a more comprehensive suite of visualization and collaboration tools, while Catalyst focuses on simplifying PyTorch training loops. wandb's cloud-based approach offers broader accessibility but may raise privacy concerns, whereas Catalyst's local logging is more privacy-friendly but potentially less convenient for remote collaboration.

Convert designs to code with AI

Introducing Visual Copilot: A new AI model to turn Figma designs to high quality code using your components.

Try Visual Copilot

README

Accelerated Deep Learning R&D

Catalyst is a PyTorch framework for Deep Learning Research and Development. It focuses on reproducibility, rapid experimentation, and codebase reuse so you can create something new rather than write yet another train loop.
Break the cycle â use the Catalyst!

Catalyst at PyTorch Ecosystem Day 2021

Catalyst at PyTorch Developer Day 2021

Getting started

pip install -U catalyst

import os
from torch import nn, optim
from torch.utils.data import DataLoader
from catalyst import dl, utils
from catalyst.contrib.datasets import MNIST

model = nn.Sequential(nn.Flatten(), nn.Linear(28 * 28, 10))
criterion = nn.CrossEntropyLoss()
optimizer = optim.Adam(model.parameters(), lr=0.02)
loaders = {
    "train": DataLoader(MNIST(os.getcwd(), train=True), batch_size=32),
    "valid": DataLoader(MNIST(os.getcwd(), train=False), batch_size=32),
}

runner = dl.SupervisedRunner(
    input_key="features", output_key="logits", target_key="targets", loss_key="loss"
)

# model training
runner.train(
    model=model,
    criterion=criterion,
    optimizer=optimizer,
    loaders=loaders,
    num_epochs=1,
    callbacks=[
        dl.AccuracyCallback(input_key="logits", target_key="targets", topk=(1, 3, 5)),
        dl.PrecisionRecallF1SupportCallback(input_key="logits", target_key="targets"),
    ],
    logdir="./logs",
    valid_loader="valid",
    valid_metric="loss",
    minimize_valid_metric=True,
    verbose=True,
)

# model evaluation
metrics = runner.evaluate_loader(
    loader=loaders["valid"],
    callbacks=[dl.AccuracyCallback(input_key="logits", target_key="targets", topk=(1, 3, 5))],
)

# model inference
for prediction in runner.predict_loader(loader=loaders["valid"]):
    assert prediction["logits"].detach().cpu().numpy().shape[-1] == 10

# model post-processing
model = runner.model.cpu()
batch = next(iter(loaders["valid"]))[0]
utils.trace_model(model=model, batch=batch)
utils.quantize_model(model=model)
utils.prune_model(model=model, pruning_fn="l1_unstructured", amount=0.8)
utils.onnx_export(model=model, batch=batch, file="./logs/mnist.onnx", verbose=True)

Step-by-step Guide

Start with Catalyst â A PyTorch Framework for Accelerated Deep Learning R&D introduction.
Try notebook tutorials or check minimal examples for first deep dive.
Read blog posts with use-cases and guides.
Learn machine learning with our "Deep Learning with Catalyst" course.
And finally, join our slack if you want to chat with the team and contributors.

Getting started
- Step-by-step Guide
Table of Contents
Overview
Community

Overview

Catalyst helps you implement compact but full-featured Deep Learning pipelines with just a few lines of code. You get a training loop with metrics, early-stopping, model checkpointing, and other features without the boilerplate.

Installation

Generic installation:

pip install -U catalyst

Specialized versions, extra requirements might apply

pip install catalyst[ml]         # installs ML-based Catalyst
pip install catalyst[cv]         # installs CV-based Catalyst
# master version installation
pip install git+https://github.com/catalyst-team/catalyst@master --upgrade
# all available extensions are listed here:
# https://github.com/catalyst-team/catalyst/blob/master/setup.py

Catalyst is compatible with: Python 3.7+. PyTorch 1.4+.
Tested on Ubuntu 16.04/18.04/20.04, macOS 10.15, Windows 10, and Windows Subsystem for Linux.

Documentation

master
22.02
2021 edition
2020 edition
- 20.12
- 20.11
- 20.10
- 20.09
- 20.08.2
- 20.07 (dev blog: 20.07 release)
- 20.06
- 20.05, 20.05.1
- 20.04, 20.04.1, 20.04.2

Minimal Examples

Introduction tutorial "Customizing what happens in train"
Demo with customization examples
Reinforcement Learning with Catalyst
And more

CustomRunner â PyTorch for-loop decomposition

import os
from torch import nn, optim
from torch.nn import functional as F
from torch.utils.data import DataLoader
from catalyst import dl, metrics
from catalyst.contrib.datasets import MNIST

model = nn.Sequential(nn.Flatten(), nn.Linear(28 * 28, 10))
optimizer = optim.Adam(model.parameters(), lr=0.02)

train_data = MNIST(os.getcwd(), train=True)
valid_data = MNIST(os.getcwd(), train=False)
loaders = {
    "train": DataLoader(train_data, batch_size=32),
    "valid": DataLoader(valid_data, batch_size=32),
}

class CustomRunner(dl.Runner):
    def predict_batch(self, batch):
        # model inference step
        return self.model(batch[0].to(self.engine.device))

    def on_loader_start(self, runner):
        super().on_loader_start(runner)
        self.meters = {
            key: metrics.AdditiveMetric(compute_on_call=False)
            for key in ["loss", "accuracy01", "accuracy03"]
        }

    def handle_batch(self, batch):
        # model train/valid step
        # unpack the batch
        x, y = batch
        # run model forward pass
        logits = self.model(x)
        # compute the loss
        loss = F.cross_entropy(logits, y)
        # compute the metrics
        accuracy01, accuracy03 = metrics.accuracy(logits, y, topk=(1, 3))
        # log metrics
        self.batch_metrics.update(
            {"loss": loss, "accuracy01": accuracy01, "accuracy03": accuracy03}
        )
        for key in ["loss", "accuracy01", "accuracy03"]:
            self.meters[key].update(self.batch_metrics[key].item(), self.batch_size)
        # run model backward pass
        if self.is_train_loader:
            self.engine.backward(loss)
            self.optimizer.step()
            self.optimizer.zero_grad()

    def on_loader_end(self, runner):
        for key in ["loss", "accuracy01", "accuracy03"]:
            self.loader_metrics[key] = self.meters[key].compute()[0]
        super().on_loader_end(runner)

runner = CustomRunner()
# model training
runner.train(
    model=model,
    optimizer=optimizer,
    loaders=loaders,
    logdir="./logs",
    num_epochs=5,
    verbose=True,
    valid_loader="valid",
    valid_metric="loss",
    minimize_valid_metric=True,
)
# model inference
for logits in runner.predict_loader(loader=loaders["valid"]):
    assert logits.detach().cpu().numpy().shape[-1] == 10

ML - linear regression

import torch
from torch.utils.data import DataLoader, TensorDataset
from catalyst import dl

# data
num_samples, num_features = int(1e4), int(1e1)
X, y = torch.rand(num_samples, num_features), torch.rand(num_samples)
dataset = TensorDataset(X, y)
loader = DataLoader(dataset, batch_size=32, num_workers=1)
loaders = {"train": loader, "valid": loader}

# model, criterion, optimizer, scheduler
model = torch.nn.Linear(num_features, 1)
criterion = torch.nn.MSELoss()
optimizer = torch.optim.Adam(model.parameters())
scheduler = torch.optim.lr_scheduler.MultiStepLR(optimizer, [3, 6])

# model training
runner = dl.SupervisedRunner()
runner.train(
    model=model,
    criterion=criterion,
    optimizer=optimizer,
    scheduler=scheduler,
    loaders=loaders,
    logdir="./logdir",
    valid_loader="valid",
    valid_metric="loss",
    minimize_valid_metric=True,
    num_epochs=8,
    verbose=True,
)

ML - multiclass classification

import torch
from torch.utils.data import DataLoader, TensorDataset
from catalyst import dl

# sample data
num_samples, num_features, num_classes = int(1e4), int(1e1), 4
X = torch.rand(num_samples, num_features)
y = (torch.rand(num_samples,) * num_classes).to(torch.int64)

# pytorch loaders
dataset = TensorDataset(X, y)
loader = DataLoader(dataset, batch_size=32, num_workers=1)
loaders = {"train": loader, "valid": loader}

# model, criterion, optimizer, scheduler
model = torch.nn.Linear(num_features, num_classes)
criterion = torch.nn.CrossEntropyLoss()
optimizer = torch.optim.Adam(model.parameters())
scheduler = torch.optim.lr_scheduler.MultiStepLR(optimizer, [2])

# model training
runner = dl.SupervisedRunner(
    input_key="features", output_key="logits", target_key="targets", loss_key="loss"
)
runner.train(
    model=model,
    criterion=criterion,
    optimizer=optimizer,
    scheduler=scheduler,
    loaders=loaders,
    logdir="./logdir",
    num_epochs=3,
    valid_loader="valid",
    valid_metric="accuracy03",
    minimize_valid_metric=False,
    verbose=True,
    callbacks=[
        dl.AccuracyCallback(input_key="logits", target_key="targets", num_classes=num_classes),
        # uncomment for extra metrics:
        # dl.PrecisionRecallF1SupportCallback(
        #     input_key="logits", target_key="targets", num_classes=num_classes
        # ),
        # dl.AUCCallback(input_key="logits", target_key="targets"),
        # catalyst[ml] required ``pip install catalyst[ml]``
        # dl.ConfusionMatrixCallback(
        #     input_key="logits", target_key="targets", num_classes=num_classes
        # ),
    ],
)

ML - multilabel classification

import torch
from torch.utils.data import DataLoader, TensorDataset
from catalyst import dl

# sample data
num_samples, num_features, num_classes = int(1e4), int(1e1), 4
X = torch.rand(num_samples, num_features)
y = (torch.rand(num_samples, num_classes) > 0.5).to(torch.float32)

# pytorch loaders
dataset = TensorDataset(X, y)
loader = DataLoader(dataset, batch_size=32, num_workers=1)
loaders = {"train": loader, "valid": loader}

# model, criterion, optimizer, scheduler
model = torch.nn.Linear(num_features, num_classes)
criterion = torch.nn.BCEWithLogitsLoss()
optimizer = torch.optim.Adam(model.parameters())
scheduler = torch.optim.lr_scheduler.MultiStepLR(optimizer, [2])

# model training
runner = dl.SupervisedRunner(
    input_key="features", output_key="logits", target_key="targets", loss_key="loss"
)
runner.train(
    model=model,
    criterion=criterion,
    optimizer=optimizer,
    scheduler=scheduler,
    loaders=loaders,
    logdir="./logdir",
    num_epochs=3,
    valid_loader="valid",
    valid_metric="accuracy01",
    minimize_valid_metric=False,
    verbose=True,
    callbacks=[
        dl.BatchTransformCallback(
            transform=torch.sigmoid,
            scope="on_batch_end",
            input_key="logits",
            output_key="scores"
        ),
        dl.AUCCallback(input_key="scores", target_key="targets"),
        # uncomment for extra metrics:
        # dl.MultilabelAccuracyCallback(input_key="scores", target_key="targets", threshold=0.5),
        # dl.MultilabelPrecisionRecallF1SupportCallback(
        #     input_key="scores", target_key="targets", threshold=0.5
        # ),
    ]
)

ML - multihead classification

import torch
from torch import nn, optim
from torch.utils.data import DataLoader, TensorDataset
from catalyst import dl

# sample data
num_samples, num_features, num_classes1, num_classes2 = int(1e4), int(1e1), 4, 10
X = torch.rand(num_samples, num_features)
y1 = (torch.rand(num_samples,) * num_classes1).to(torch.int64)
y2 = (torch.rand(num_samples,) * num_classes2).to(torch.int64)

# pytorch loaders
dataset = TensorDataset(X, y1, y2)
loader = DataLoader(dataset, batch_size=32, num_workers=1)
loaders = {"train": loader, "valid": loader}

class CustomModule(nn.Module):
    def __init__(self, in_features: int, out_features1: int, out_features2: int):
        super().__init__()
        self.shared = nn.Linear(in_features, 128)
        self.head1 = nn.Linear(128, out_features1)
        self.head2 = nn.Linear(128, out_features2)

    def forward(self, x):
        x = self.shared(x)
        y1 = self.head1(x)
        y2 = self.head2(x)
        return y1, y2

# model, criterion, optimizer, scheduler
model = CustomModule(num_features, num_classes1, num_classes2)
criterion = nn.CrossEntropyLoss()
optimizer = optim.Adam(model.parameters())
scheduler = optim.lr_scheduler.MultiStepLR(optimizer, [2])

class CustomRunner(dl.Runner):
    def handle_batch(self, batch):
        x, y1, y2 = batch
        y1_hat, y2_hat = self.model(x)
        self.batch = {
            "features": x,
            "logits1": y1_hat,
            "logits2": y2_hat,
            "targets1": y1,
            "targets2": y2,
        }

# model training
runner = CustomRunner()
runner.train(
    model=model,
    criterion=criterion,
    optimizer=optimizer,
    scheduler=scheduler,
    loaders=loaders,
    num_epochs=3,
    verbose=True,
    callbacks=[
        dl.CriterionCallback(metric_key="loss1", input_key="logits1", target_key="targets1"),
        dl.CriterionCallback(metric_key="loss2", input_key="logits2", target_key="targets2"),
        dl.MetricAggregationCallback(metric_key="loss", metrics=["loss1", "loss2"], mode="mean"),
        dl.BackwardCallback(metric_key="loss"),
        dl.OptimizerCallback(metric_key="loss"),
        dl.SchedulerCallback(),
        dl.AccuracyCallback(
            input_key="logits1", target_key="targets1", num_classes=num_classes1, prefix="one_"
        ),
        dl.AccuracyCallback(
            input_key="logits2", target_key="targets2", num_classes=num_classes2, prefix="two_"
        ),
        # catalyst[ml] required ``pip install catalyst[ml]``
        # dl.ConfusionMatrixCallback(
        #     input_key="logits1", target_key="targets1", num_classes=num_classes1, prefix="one_cm"
        # ),
        # dl.ConfusionMatrixCallback(
        #     input_key="logits2", target_key="targets2", num_classes=num_classes2, prefix="two_cm"
        # ),
        dl.CheckpointCallback(
            logdir="./logs/one",
            loader_key="valid", metric_key="one_accuracy01", minimize=False, topk=1
        ),
        dl.CheckpointCallback(
            logdir="./logs/two",
            loader_key="valid", metric_key="two_accuracy03", minimize=False, topk=3
        ),
    ],
    loggers={"console": dl.ConsoleLogger(), "tb": dl.TensorboardLogger("./logs/tb")},
)

ML â RecSys

import torch
from torch.utils.data import DataLoader, TensorDataset
from catalyst import dl

# sample data
num_users, num_features, num_items = int(1e4), int(1e1), 10
X = torch.rand(num_users, num_features)
y = (torch.rand(num_users, num_items) > 0.5).to(torch.float32)

# pytorch loaders
dataset = TensorDataset(X, y)
loader = DataLoader(dataset, batch_size=32, num_workers=1)
loaders = {"train": loader, "valid": loader}

# model, criterion, optimizer, scheduler
model = torch.nn.Linear(num_features, num_items)
criterion = torch.nn.BCEWithLogitsLoss()
optimizer = torch.optim.Adam(model.parameters())
scheduler = torch.optim.lr_scheduler.MultiStepLR(optimizer, [2])

# model training
runner = dl.SupervisedRunner(
    input_key="features", output_key="logits", target_key="targets", loss_key="loss"
)
runner.train(
    model=model,
    criterion=criterion,
    optimizer=optimizer,
    scheduler=scheduler,
    loaders=loaders,
    num_epochs=3,
    verbose=True,
    callbacks=[
        dl.BatchTransformCallback(
            transform=torch.sigmoid,
            scope="on_batch_end",
            input_key="logits",
            output_key="scores"
        ),
        dl.CriterionCallback(input_key="logits", target_key="targets", metric_key="loss"),
        # uncomment for extra metrics:
        # dl.AUCCallback(input_key="scores", target_key="targets"),
        # dl.HitrateCallback(input_key="scores", target_key="targets", topk=(1, 3, 5)),
        # dl.MRRCallback(input_key="scores", target_key="targets", topk=(1, 3, 5)),
        # dl.MAPCallback(input_key="scores", target_key="targets", topk=(1, 3, 5)),
        # dl.NDCGCallback(input_key="scores", target_key="targets", topk=(1, 3, 5)),
        dl.BackwardCallback(metric_key="loss"),
        dl.OptimizerCallback(metric_key="loss"),
        dl.SchedulerCallback(),
        dl.CheckpointCallback(
            logdir="./logs", loader_key="valid", metric_key="loss", minimize=True
        ),
    ]
)

CV - MNIST classification

import os
from torch import nn, optim
from torch.utils.data import DataLoader
from catalyst import dl
from catalyst.contrib.datasets import MNIST

model = nn.Sequential(nn.Flatten(), nn.Linear(28 * 28, 10))
criterion = nn.CrossEntropyLoss()
optimizer = optim.Adam(model.parameters(), lr=0.02)

train_data = MNIST(os.getcwd(), train=True)
valid_data = MNIST(os.getcwd(), train=False)
loaders = {
    "train": DataLoader(train_data, batch_size=32),
    "valid": DataLoader(valid_data, batch_size=32),
}

runner = dl.SupervisedRunner()
# model training
runner.train(
    model=model,
    criterion=criterion,
    optimizer=optimizer,
    loaders=loaders,
    num_epochs=1,
    logdir="./logs",
    valid_loader="valid",
    valid_metric="loss",
    minimize_valid_metric=True,
    verbose=True,
# uncomment for extra metrics:
#     callbacks=[
#         dl.AccuracyCallback(input_key="logits", target_key="targets", num_classes=10),
#         dl.PrecisionRecallF1SupportCallback(
#             input_key="logits", target_key="targets", num_classes=10
#         ),
#         dl.AUCCallback(input_key="logits", target_key="targets"),
#         # catalyst[ml] required ``pip install catalyst[ml]``
#         dl.ConfusionMatrixCallback(
#             input_key="logits", target_key="targets", num_classes=num_classes
#         ),
#     ]
)

CV - MNIST segmentation

import os
import torch
from torch import nn
from torch.utils.data import DataLoader
from catalyst import dl
from catalyst.contrib.datasets import MNIST
from catalyst.contrib.losses import IoULoss


model = nn.Sequential(
    nn.Conv2d(1, 1, 3, 1, 1), nn.ReLU(),
    nn.Conv2d(1, 1, 3, 1, 1), nn.Sigmoid(),
)
criterion = IoULoss()
optimizer = torch.optim.Adam(model.parameters(), lr=0.02)

train_data = MNIST(os.getcwd(), train=True)
valid_data = MNIST(os.getcwd(), train=False)
loaders = {
    "train": DataLoader(train_data, batch_size=32),
    "valid": DataLoader(valid_data, batch_size=32),
}

class CustomRunner(dl.SupervisedRunner):
    def handle_batch(self, batch):
        x = batch[self._input_key]
        x_noise = (x + torch.rand_like(x)).clamp_(0, 1)
        x_ = self.model(x_noise)
        self.batch = {self._input_key: x, self._output_key: x_, self._target_key: x}

runner = CustomRunner(
    input_key="features", output_key="scores", target_key="targets", loss_key="loss"
)
# model training
runner.train(
    model=model,
    criterion=criterion,
    optimizer=optimizer,
    loaders=loaders,
    num_epochs=1,
    callbacks=[
        dl.IOUCallback(input_key="scores", target_key="targets"),
        dl.DiceCallback(input_key="scores", target_key="targets"),
        dl.TrevskyCallback(input_key="scores", target_key="targets", alpha=0.2),
    ],
    logdir="./logdir",
    valid_loader="valid",
    valid_metric="loss",
    minimize_valid_metric=True,
    verbose=True,
)

CV - MNIST metric learning

import os
from torch.optim import Adam
from torch.utils.data import DataLoader
from catalyst import dl
from catalyst.contrib.data import HardTripletsSampler
from catalyst.contrib.datasets import MnistMLDataset, MnistQGDataset
from catalyst.contrib.losses import TripletMarginLossWithSampler
from catalyst.contrib.models import MnistSimpleNet
from catalyst.data.sampler import BatchBalanceClassSampler


# 1. train and valid loaders
train_dataset = MnistMLDataset(root=os.getcwd())
sampler = BatchBalanceClassSampler(
    labels=train_dataset.get_labels(), num_classes=5, num_samples=10, num_batches=10
)
train_loader = DataLoader(dataset=train_dataset, batch_sampler=sampler)

valid_dataset = MnistQGDataset(root=os.getcwd(), gallery_fraq=0.2)
valid_loader = DataLoader(dataset=valid_dataset, batch_size=1024)

# 2. model and optimizer
model = MnistSimpleNet(out_features=16)
optimizer = Adam(model.parameters(), lr=0.001)

# 3. criterion with triplets sampling
sampler_inbatch = HardTripletsSampler(norm_required=False)
criterion = TripletMarginLossWithSampler(margin=0.5, sampler_inbatch=sampler_inbatch)

# 4. training with catalyst Runner
class CustomRunner(dl.SupervisedRunner):
    def handle_batch(self, batch) -> None:
        if self.is_train_loader:
            images, targets = batch["features"].float(), batch["targets"].long()
            features = self.model(images)
            self.batch = {"embeddings": features, "targets": targets,}
        else:
            images, targets, is_query = \
                batch["features"].float(), batch["targets"].long(), batch["is_query"].bool()
            features = self.model(images)
            self.batch = {"embeddings": features, "targets": targets, "is_query": is_query}

callbacks = [
    dl.ControlFlowCallbackWrapper(
        dl.CriterionCallback(input_key="embeddings", target_key="targets", metric_key="loss"),
        loaders="train",
    ),
    dl.ControlFlowCallbackWrapper(
        dl.CMCScoreCallback(
            embeddings_key="embeddings",
            labels_key="targets",
            is_query_key="is_query",
            topk=[1],
        ),
        loaders="valid",
    ),
    dl.PeriodicLoaderCallback(
        valid_loader_key="valid", valid_metric_key="cmc01", minimize=False, valid=2
    ),
]

runner = CustomRunner(input_key="features", output_key="embeddings")
runner.train(
    model=model,
    criterion=criterion,
    optimizer=optimizer,
    callbacks=callbacks,
    loaders={"train": train_loader, "valid": valid_loader},
    verbose=False,
    logdir="./logs",
    valid_loader="valid",
    valid_metric="cmc01",
    minimize_valid_metric=False,
    num_epochs=10,
)

CV - MNIST GAN

import os
import torch
from torch import nn
from torch.utils.data import DataLoader
from catalyst import dl
from catalyst.contrib.datasets import MNIST
from catalyst.contrib.layers import GlobalMaxPool2d, Lambda

latent_dim = 128
generator = nn.Sequential(
    # We want to generate 128 coefficients to reshape into a 7x7x128 map
    nn.Linear(128, 128 * 7 * 7),
    nn.LeakyReLU(0.2, inplace=True),
    Lambda(lambda x: x.view(x.size(0), 128, 7, 7)),
    nn.ConvTranspose2d(128, 128, (4, 4), stride=(2, 2), padding=1),
    nn.LeakyReLU(0.2, inplace=True),
    nn.ConvTranspose2d(128, 128, (4, 4), stride=(2, 2), padding=1),
    nn.LeakyReLU(0.2, inplace=True),
    nn.Conv2d(128, 1, (7, 7), padding=3),
    nn.Sigmoid(),
)
discriminator = nn.Sequential(
    nn.Conv2d(1, 64, (3, 3), stride=(2, 2), padding=1),
    nn.LeakyReLU(0.2, inplace=True),
    nn.Conv2d(64, 128, (3, 3), stride=(2, 2), padding=1),
    nn.LeakyReLU(0.2, inplace=True),
    GlobalMaxPool2d(),
    nn.Flatten(),
    nn.Linear(128, 1),
)

model = nn.ModuleDict({"generator": generator, "discriminator": discriminator})
criterion = {"generator": nn.BCEWithLogitsLoss(), "discriminator": nn.BCEWithLogitsLoss()}
optimizer = {
    "generator": torch.optim.Adam(generator.parameters(), lr=0.0003, betas=(0.5, 0.999)),
    "discriminator": torch.optim.Adam(discriminator.parameters(), lr=0.0003, betas=(0.5, 0.999)),
}
train_data = MNIST(os.getcwd(), train=False)
loaders = {"train": DataLoader(train_data, batch_size=32)}

class CustomRunner(dl.Runner):
    def predict_batch(self, batch):
        batch_size = 1
        # Sample random points in the latent space
        random_latent_vectors = torch.randn(batch_size, latent_dim).to(self.engine.device)
        # Decode them to fake images
        generated_images = self.model["generator"](random_latent_vectors).detach()
        return generated_images

    def handle_batch(self, batch):
        real_images, _ = batch
        batch_size = real_images.shape[0]

        # Sample random points in the latent space
        random_latent_vectors = torch.randn(batch_size, latent_dim).to(self.engine.device)

        # Decode them to fake images
        generated_images = self.model["generator"](random_latent_vectors).detach()
        # Combine them with real images
        combined_images = torch.cat([generated_images, real_images])

        # Assemble labels discriminating real from fake images
        labels = \
            torch.cat([torch.ones((batch_size, 1)), torch.zeros((batch_size, 1))]).to(self.engine.device)
        # Add random noise to the labels - important trick!
        labels += 0.05 * torch.rand(labels.shape).to(self.engine.device)

        # Discriminator forward
        combined_predictions = self.model["discriminator"](combined_images)

        # Sample random points in the latent space
        random_latent_vectors = torch.randn(batch_size, latent_dim).to(self.engine.device)
        # Assemble labels that say "all real images"
        misleading_labels = torch.zeros((batch_size, 1)).to(self.engine.device)

        # Generator forward
        generated_images = self.model["generator"](random_latent_vectors)
        generated_predictions = self.model["discriminator"](generated_images)

        self.batch = {
            "combined_predictions": combined_predictions,
            "labels": labels,
            "generated_predictions": generated_predictions,
            "misleading_labels": misleading_labels,
        }


runner = CustomRunner()
runner.train(
    model=model,
    criterion=criterion,
    optimizer=optimizer,
    loaders=loaders,
    callbacks=[
        dl.CriterionCallback(
            input_key="combined_predictions",
            target_key="labels",
            metric_key="loss_discriminator",
            criterion_key="discriminator",
        ),
        dl.BackwardCallback(metric_key="loss_discriminator"),
        dl.OptimizerCallback(
            optimizer_key="discriminator",
            metric_key="loss_discriminator",
        ),
        dl.CriterionCallback(
            input_key="generated_predictions",
            target_key="misleading_labels",
            metric_key="loss_generator",
            criterion_key="generator",
        ),
        dl.BackwardCallback(metric_key="loss_generator"),
        dl.OptimizerCallback(
            optimizer_key="generator",
            metric_key="loss_generator",
        ),
    ],
    valid_loader="train",
    valid_metric="loss_generator",
    minimize_valid_metric=True,
    num_epochs=20,
    verbose=True,
    logdir="./logs_gan",
)

# visualization (matplotlib required):
# import matplotlib.pyplot as plt
# %matplotlib inline
# plt.imshow(runner.predict_batch(None)[0, 0].cpu().numpy())

CV - MNIST VAE

import os
import torch
from torch import nn, optim
from torch.nn import functional as F
from torch.utils.data import DataLoader
from catalyst import dl, metrics
from catalyst.contrib.datasets import MNIST

LOG_SCALE_MAX = 2
LOG_SCALE_MIN = -10

def normal_sample(loc, log_scale):
    scale = torch.exp(0.5 * log_scale)
    return loc + scale * torch.randn_like(scale)

class VAE(nn.Module):
    def __init__(self, in_features, hid_features):
        super().__init__()
        self.hid_features = hid_features
        self.encoder = nn.Linear(in_features, hid_features * 2)
        self.decoder = nn.Sequential(nn.Linear(hid_features, in_features), nn.Sigmoid())

    def forward(self, x, deterministic=False):
        z = self.encoder(x)
        bs, z_dim = z.shape

        loc, log_scale = z[:, : z_dim // 2], z[:, z_dim // 2 :]
        log_scale = torch.clamp(log_scale, LOG_SCALE_MIN, LOG_SCALE_MAX)

        z_ = loc if deterministic else normal_sample(loc, log_scale)
        z_ = z_.view(bs, -1)
        x_ = self.decoder(z_)

        return x_, loc, log_scale

class CustomRunner(dl.IRunner):
    def __init__(self, hid_features, logdir, engine):
        super().__init__()
        self.hid_features = hid_features
        self._logdir = logdir
        self._engine = engine

    def get_engine(self):
        return self._engine

    def get_loggers(self):
        return {
            "console": dl.ConsoleLogger(),
            "csv": dl.CSVLogger(logdir=self._logdir),
            "tensorboard": dl.TensorboardLogger(logdir=self._logdir),
        }

    @property
    def num_epochs(self) -> int:
        return 1

    def get_loaders(self):
        loaders = {
            "train": DataLoader(MNIST(os.getcwd(), train=False), batch_size=32),
            "valid": DataLoader(MNIST(os.getcwd(), train=False), batch_size=32),
        }
        return loaders

    def get_model(self):
        model = self.model if self.model is not None else VAE(28 * 28, self.hid_features)
        return model

    def get_optimizer(self, model):
        return optim.Adam(model.parameters(), lr=0.02)

    def get_callbacks(self):
        return {
            "backward": dl.BackwardCallback(metric_key="loss"),
            "optimizer": dl.OptimizerCallback(metric_key="loss"),
            "checkpoint": dl.CheckpointCallback(
                self._logdir,
                loader_key="valid",
                metric_key="loss",
                minimize=True,
                topk=3,
            ),
        }

    def on_loader_start(self, runner):
        super().on_loader_start(runner)
        self.meters = {
            key: metrics.AdditiveMetric(compute_on_call=False)
            for key in ["loss_ae", "loss_kld", "loss"]
        }

    def handle_batch(self, batch):
        x, _ = batch
        x = x.view(x.size(0), -1)
        x_, loc, log_scale = self.model(x, deterministic=not self.is_train_loader)

        loss_ae = F.mse_loss(x_, x)
        loss_kld = (
            -0.5 * torch.sum(1 + log_scale - loc.pow(2) - log_scale.exp(), dim=1)
        ).mean()
        loss = loss_ae + loss_kld * 0.01

        self.batch_metrics = {"loss_ae": loss_ae, "loss_kld": loss_kld, "loss": loss}
        for key in ["loss_ae", "loss_kld", "loss"]:
            self.meters[key].update(self.batch_metrics[key].item(), self.batch_size)

    def on_loader_end(self, runner):
        for key in ["loss_ae", "loss_kld", "loss"]:
            self.loader_metrics[key] = self.meters[key].compute()[0]
        super().on_loader_end(runner)

    def predict_batch(self, batch):
        random_latent_vectors = torch.randn(1, self.hid_features).to(self.engine.device)
        generated_images = self.model.decoder(random_latent_vectors).detach()
        return generated_images

runner = CustomRunner(128, "./logs", dl.CPUEngine())
runner.run()
# visualization (matplotlib required):
# import matplotlib.pyplot as plt
# %matplotlib inline
# plt.imshow(runner.predict_batch(None)[0].cpu().numpy().reshape(28, 28))

AutoML - hyperparameters optimization with Optuna

import os
import optuna
import torch
from torch import nn
from torch.utils.data import DataLoader
from catalyst import dl
from catalyst.contrib.datasets import MNIST


def objective(trial):
    lr = trial.suggest_loguniform("lr", 1e-3, 1e-1)
    num_hidden = int(trial.suggest_loguniform("num_hidden", 32, 128))

    train_data = MNIST(os.getcwd(), train=True)
    valid_data = MNIST(os.getcwd(), train=False)
    loaders = {
        "train": DataLoader(train_data, batch_size=32),
        "valid": DataLoader(valid_data, batch_size=32),
    }
    model = nn.Sequential(
        nn.Flatten(), nn.Linear(784, num_hidden), nn.ReLU(), nn.Linear(num_hidden, 10)
    )
    optimizer = torch.optim.Adam(model.parameters(), lr=lr)
    criterion = nn.CrossEntropyLoss()

    runner = dl.SupervisedRunner(input_key="features", output_key="logits", target_key="targets")
    runner.train(
        model=model,
        criterion=criterion,
        optimizer=optimizer,
        loaders=loaders,
        callbacks={
            "accuracy": dl.AccuracyCallback(
                input_key="logits", target_key="targets", num_classes=10
            ),
            # catalyst[optuna] required ``pip install catalyst[optuna]``
            "optuna": dl.OptunaPruningCallback(
                loader_key="valid", metric_key="accuracy01", minimize=False, trial=trial
            ),
        },
        num_epochs=3,
    )
    score = trial.best_score
    return score

study = optuna.create_study(
    direction="maximize",
    pruner=optuna.pruners.MedianPruner(
        n_startup_trials=1, n_warmup_steps=0, interval_steps=1
    ),
)
study.optimize(objective, n_trials=3, timeout=300)
print(study.best_value, study.best_params)

Config API - minimal example

runner:
  _target_: catalyst.runners.SupervisedRunner
  model:
    _var_: model
    _target_: torch.nn.Sequential
    args:
      - _target_: torch.nn.Flatten
      - _target_: torch.nn.Linear
        in_features: 784  # 28 * 28
        out_features: 10
  input_key: features
  output_key: &output_key logits
  target_key: &target_key targets
  loss_key: &loss_key loss

run:
  # â stage 1
  - _call_: train  # runner.train(...)

    criterion:
      _target_: torch.nn.CrossEntropyLoss

    optimizer:
      _target_: torch.optim.Adam
      params:  # model.parameters()
        _var_: model.parameters
      lr: 0.02

    loaders:
      train:
        _target_: torch.utils.data.DataLoader
        dataset:
          _target_: catalyst.contrib.datasets.MNIST
          root: data
          train: y
        batch_size: 32

      &valid_loader_key valid:
        &valid_loader
        _target_: torch.utils.data.DataLoader
        dataset:
          _target_: catalyst.contrib.datasets.MNIST
          root: data
          train: n
        batch_size: 32

    callbacks:
      - &accuracy_metric
        _target_: catalyst.callbacks.AccuracyCallback
        input_key: *output_key
        target_key: *target_key
        topk: [1,3,5]
      - _target_: catalyst.callbacks.PrecisionRecallF1SupportCallback
        input_key: *output_key
        target_key: *target_key

    num_epochs: 1
    logdir: logs
    valid_loader: *valid_loader_key
    valid_metric: *loss_key
    minimize_valid_metric: y
    verbose: y

  # â stage 2
  - _call_: evaluate_loader  # runner.evaluate_loader(...)
    loader: *valid_loader
    callbacks:
      - *accuracy_metric

catalyst-run --config example.yaml

Tests

All Catalyst code, features, and pipelines are fully tested. We also have our own catalyst-codestyle and a corresponding pre-commit hook. During testing, we train a variety of different models: image classification, image segmentation, text classification, GANs, and much more. We then compare their convergence metrics in order to verify the correctness of the training procedure and its reproducibility. As a result, Catalyst provides fully tested and reproducible best practices for your deep learning research and development.

Blog Posts

Talks

Community

Accelerated with Catalyst

Research Papers

Blog Posts

Competitions

Kaggle Quick, Draw! Doodle Recognition Challenge - 11th place
Catalyst.RL - NeurIPS 2018: AI for Prosthetics Challenge â 3rd place
Kaggle Google Landmark 2019 - 30th place
iMet Collection 2019 - FGVC6 - 24th place
ID R&D Anti-spoofing Challenge - 14th place
NeurIPS 2019: Recursion Cellular Image Classification - 4th place
MICCAI 2019: Automatic Structure Segmentation for Radiotherapy Planning Challenge 2019
- 3rd place solution for Task 3: Organ-at-risk segmentation from chest CT scans
- and 4th place solution for Task 4: Gross Target Volume segmentation of lung cancer
Kaggle Seversteal steel detection - 5th place
RSNA Intracranial Hemorrhage Detection - 5th place
APTOS 2019 Blindness Detection â 7th place
Catalyst.RL - NeurIPS 2019: Learn to Move - Walk Around â 2nd place
xView2 Damage Assessment Challenge - 3rd place

Toolkits

Catalyst.RL â A Distributed Framework for Reproducible RL Research by Scitator
Catalyst.Classification - Comprehensive classification pipeline with Pseudo-Labeling by Bagxi and Pdanilov
Catalyst.Segmentation - Segmentation pipelines - binary, semantic and instance, by Bagxi
Catalyst.Detection - Anchor-free detection pipeline by Avi2011class and TezRomacH
Catalyst.GAN - Reproducible GANs pipelines by Asmekal
Catalyst.Neuro - Brain image analysis project, in collaboration with TReNDS Center
MLComp â Distributed DAG framework for machine learning with UI by Lightforever
Pytorch toolbelt - PyTorch extensions for fast R&D prototyping and Kaggle farming by BloodAxe
Helper functions - An assorted collection of helper functions by Ternaus
BERT Distillation with Catalyst by elephantmipt

Other

CamVid Segmentation Example - Example of semantic segmentation for CamVid dataset
Notebook API tutorial for segmentation in Understanding Clouds from Satellite Images Competition
Catalyst.RL - NeurIPS 2019: Learn to Move - Walk Around â starter kit
Catalyst.RL - NeurIPS 2019: Animal-AI Olympics - starter kit
Inria Segmentation Example - An example of training segmentation model for Inria Sattelite Segmentation Challenge
iglovikov_segmentation - Semantic segmentation pipeline using Catalyst
Logging Catalyst Runs to Comet - An example of how to log metrics, hyperparameters and more from Catalyst runs to Comet

See other projects at the GitHub dependency graph.

If your project implements a paper, a notable use-case/tutorial, or a Kaggle competition solution, or if your code simply presents interesting results and uses Catalyst, we would be happy to add your project to the list above! Do not hesitate to send us a PR with a brief description of the project similar to the above.

Contribution Guide

We appreciate all contributions. If you are planning to contribute back bug-fixes, there is no need to run that by us; just send a PR. If you plan to contribute new features, new utility functions, or extensions, please open an issue first and discuss it with us.

Please see the Contribution Guide for more information.
By participating in this project, you agree to abide by its Code of Conduct.

User Feedback

We've created feedback@catalyst-team.com as an additional channel for user feedback.

If you like the project and want to thank us, this is the right place.
If you would like to start a collaboration between your team and Catalyst team to improve Deep Learning R&D, you are always welcome.
If you don't like Github Issues and prefer email, feel free to email us.
Finally, if you do not like something, please, share it with us, and we can see how to improve it.

We appreciate any type of feedback. Thank you!

Acknowledgments

Since the beginning of the Ð¡atalyst development, a lot of people have influenced it in a lot of different ways.

Catalyst.Team

Catalyst.Contributors

Trusted by

Awecom
Researchers at the Center for Translational Research in Neuroimaging and Data Science (TReNDS)
Deep Learning School
Researchers at Emory University
Evil Martians
Researchers at the Georgia Institute of Technology
Researchers at Georgia State University
Helios
HPCD Lab
iFarm
Kinoplan
Researchers at the Moscow Institute of Physics and Technology
Neuromation
Poteha Labs
Provectus
Researchers at the Skolkovo Institute of Science and Technology
SoftConstruct
Researchers at Tinkoff
Researchers at Yandex.Research

Citation

Please use this bibtex if you want to cite this repository in your publications:

@misc{catalyst,
    author = {Kolesnikov, Sergey},
    title = {Catalyst - Accelerated deep learning R&D},
    year = {2018},
    publisher = {GitHub},
    journal = {GitHub repository},
    howpublished = {\url{https://github.com/catalyst-team/catalyst}},
}

Top Related Projects

Convert designs to code with AI

Introducing Visual Copilot: A new AI model to turn Figma designs to high quality code using your components.

Try Visual Copilot

Top Related Projects

Quick Overview

Pros

Cons

Code Examples

Getting Started

Competitor Comparisons

Pros of Ignite

Cons of Ignite

Code Comparison

Pros of Ray

Cons of Ray

Code Comparison

Pros of DeepSpeed

Cons of DeepSpeed

Code Comparison

Pros of Horovod

Cons of Horovod

Code Comparison

Pros of MLflow

Cons of MLflow

Code Comparison

Pros of wandb

Cons of wandb

Code Comparison

Convert designs to code with AI

README

Getting started

Step-by-step Guide

Table of Contents

Overview

Installation

Documentation

Minimal Examples

Tests

Community

Accelerated with Catalyst

Contribution Guide

User Feedback

Acknowledgments

Catalyst.Team

Catalyst.Contributors

Trusted by

Citation

Top Related Projects

Convert designs to code with AI