🔥 Cogitare - A Modern, Fast, and Modular Deep Learning and Machine Learning framework for Python

Overview

Cogitare is a Modern, Fast, and Modular Deep Learning and Machine Learning framework for Python. A friendly interface for beginners and a powerful toolset for experts.

Cogitare is built on top of PyTorch.

Documentation • Tutorials • About • Install • Quickstart • Contribution

PyPI version

1. About

It uses the best of PyTorch, Dask, NumPy, and others tools through a simple interface to train, to evaluate, to test models and more.

With Cogitare, you can use classical machine learning algorithms with high performance and develop state-of-the-art models quickly.

Check the tutorials at http://tutorials.cogitare-ai.org/

The primary objectives of Cogitare are:

  • provide an easy-to-use interface to train and evaluate models;
  • provide tools to debug and analyze the model;
  • provide implementations of state-of-the-art models (models for common tasks, ready to train and ready to use);
  • provide ready-to-use implementations of straightforward and classical models (such as LogisticRegression);
  • be compatible with models for a broad range of problems;
  • be compatible with other tools (scikit-learn, etcs);
  • keep growing with the community: accept as many new features as possible;
  • provide a friendly interface to beginners, and powerful features for experts;
  • take the best of the hardware through multi-processing and multi-threading;
  • and others.

Currently, it's a work in progress project that aims to provide a complete toolchain for machine learning and deep learning development, taking the best of cuda and multi-core processing.

2. Install

  • Install PyTorch from http://pytorch.org/

  • Install Cogitare from PIP:

    pip install cogitare
    
  • Cogitare is in active development, so it's recommended to get the latest version from GitHub. To install directly from GitHub, use:

    pip install -e git+https://github.com/cogitare-ai/cogitare#egg=cogitare
    

3. Quickstart

This is a simple tutorial to get started with Cogitare main functionalities.

In this tutorial, we will write a Convolutional Neural Network (CNN) to classify handwritten digits (MNIST).

3.1 Model

We start by defining our CNN model.

When developing a model with Cogitare, your model must extend the cogitare.Model class. This class provides the Model interface, which allows you to train and evaluate the model efficiently.

To implement a model, you must extend the cogitare.Model class and implement the forward() and loss() methods. The forward method will receive the batch. In this way, it is necessary to implement the forward pass through the network in this method, and then return the output of the net. The loss method will receive the output of the forward() and the batch received from the iterator, apply a loss function, compute and return it.

The Model interface will iterate over the dataset, and execute each batch on forward, loss, and backward.

# adapted from https://github.com/pytorch/examples/blob/master/mnist/main.py
from cogitare import Model
from cogitare import utils
from cogitare.data import DataSet, AsyncDataLoader
from cogitare.plugins import EarlyStopping
from cogitare.metrics.classification import accuracy
import cogitare

import torch.nn as nn
import torch
import torch.nn.functional as F
from torch.nn.utils import clip_grad_norm
import torch.optim as optim

from sklearn.datasets import fetch_mldata

import numpy as np

CUDA = True


cogitare.utils.set_cuda(CUDA)
class CNN(Model):
    
    def __init__(self):
        super(CNN, self).__init__()
        
        # define the model
        self.conv1 = nn.Conv2d(1, 10, kernel_size=5)
        self.conv2 = nn.Conv2d(10, 20, kernel_size=5)
        self.conv2_drop = nn.Dropout2d()
        self.fc1 = nn.Linear(320, 50)
        self.fc2 = nn.Linear(50, 10)
    
    def forward(self, batch):
        # in this sample, each batch will be a tuple containing (input_batch, expected_batch)
        # in forward in are only interested in input so that we can ignore the second item of the tuple
        input, _ = batch
        
        # batch X flat tensor -> batch X 1 channel (gray) X width X heigth
        input = input.view(32, 1, 28, 28)
        
        # pass the data in the net
        x = F.relu(F.max_pool2d(self.conv1(input), 2))
        x = F.relu(F.max_pool2d(self.conv2_drop(self.conv2(x)), 2))
        x = x.view(-1, 320)
        x = F.relu(self.fc1(x))
        x = F.dropout(x, training=self.training)
        x = self.fc2(x)

        # return the model output
        return F.log_softmax(x, dim=1)
    
    def loss(self, output, batch):
        # in this sample, each batch will be a tuple containing (input_batch, expected_batch)
        # in loss in are only interested in expected so that we can ignore the first item of the tuple
        _, expected = batch
        
        return F.nll_loss(output, expected)

The model class is simple; it only requires de forward and loss methods. By default, Cogitare will backward the loss returned by the loss() method, and optimize the model parameters. If you want to disable the Cogitare backward and optimization steps, just return None in the loss function. If you return None, you are responsible by backwarding and optimizing the parameters.

3.2 Data Loading

In this step, we will load the data from sklearn package.

mnist = fetch_mldata('MNIST original')
mnist.data = (mnist.data / 255).astype(np.float32)

Cogitare provides a toolbox to load and pre-process data for your models. In this introduction, we will use the DataSet and the AsyncDataLoader as examples.

The DataSet is responsible by iterating over multiples data iterators (in our case, we'll have two data iterators: input samples, expected samples).

# as input, the DataSet is expected a list of iterators. In our case, the first iterator is the input 
# data and the second iterator is the target data

# also, we set the batch size to 32 and enable the shuffling

# drop the last batch if its size is different of 32
data = DataSet([mnist.data, mnist.target.astype(int)], batch_size=32, shuffle=True, drop_last=True)

# then, we split our dataset into a train and into a validation sets, by a ratio of 0.8
data_train, data_validation = data.split(0.8)

Notice that Cogitare accepts any iterator as input. Instead of using our DataSet, you can use the mnist.data itself, PyTorch's data loaders, or any other input that acts as an iterator.

In some cases, we can increase the model performance by loading the data using multiples threads/processes or by pre-loading the data before being requested by the model.

With the AsyncDataLoader, we can load N batches ahead of the model execution in parallel. We present this technique in this sample because it can increase performance in a wide range of models (when the data loading or pre-processing is slower than the model execution).

def pre_process(batch):
    input, expected = batch
    
    # the data is a numpy.ndarray (loaded from sklearn), so we need to convert it to Variable
    input = utils.to_variable(input, dtype=torch.FloatTensor)  # converts to a torch Variable of LongTensor
    expected = utils.to_variable(expected, dtype=torch.LongTensor)  # converts to a torch Variable of LongTensor
    return input, expected


# we wrap our data_train and data_validation iterators over the async data loader.
# each loader will load 16 batches ahead of the model execution using 8 workers (8 threads, in this case).
# for each batch, it will be pre-processed in parallel with the preprocess function, that will load the data
# on GPU
data_train = AsyncDataLoader(data_train, buffer_size=16, mode='threaded', workers=8, on_batch_loaded=pre_process)
data_validation = AsyncDataLoader(data_validation, buffer_size=16, mode='threaded', workers=8, on_batch_loaded=pre_process)

to cache the async buffer before training, we can:

data_train.cache()
data_validation.cache()

3.3 Training

Now, we can train our model.

First, lets create the model instance and add the default plugins to watch the training status. The default plugin includes:

  • Progress bar per batch and epoch
  • Plot training and validation losses (if validation_dataset is present)
  • Log training loss
model = CNN()
model.register_default_plugins()

Besides that, we may want to add some extra plugins, such as the EarlyStopping. So, if the model is not decreasing the loss after N epochs, the training stops and the best model is used.

To add the early stopping algorithm, you can use:

early = EarlyStopping(max_tries=10, path='/tmp/model.pt')
# after 10 epochs without decreasing the loss, stop the training and the best model is saved at /tmp/model.pt

# the plugin will execute in the end of each epoch
model.register_plugin(early, 'on_end_epoch')

Also, a common technique is to clip the gradient during training. If you want to clip the grad, you can use:

model.register_plugin(lambda *args, **kw: clip_grad_norm(model.parameters(), 1.0), 'before_step')
# will execute the clip_grad_norm before each optimization step

Now, we define the optimizator, and then start the model training:

optimizer = optim.Adam(model.parameters(), lr=0.001)

if CUDA:
    model = model.cuda()
model.learn(data_train, optimizer, data_validation, max_epochs=100)
2018-02-02 20:59:23 sprawl cogitare.core.model[2443] INFO Model: 

CNN(
  (conv1): Conv2d (1, 10, kernel_size=(5, 5), stride=(1, 1))
  (conv2): Conv2d (10, 20, kernel_size=(5, 5), stride=(1, 1))
  (conv2_drop): Dropout2d(p=0.5)
  (fc1): Linear(in_features=320, out_features=50)
  (fc2): Linear(in_features=50, out_features=10)
)

2018-02-02 20:59:23 sprawl cogitare.core.model[2443] INFO Training data: 

DataSet with:
    containers: [
        TensorHolder with 1750x32 samples
	TensorHolder with 1750x32 samples
    ],
    batch size: 32


2018-02-02 20:59:23 sprawl cogitare.core.model[2443] INFO Number of trainable parameters: 21,840
2018-02-02 20:59:23 sprawl cogitare.core.model[2443] INFO Number of non-trainable parameters: 0
2018-02-02 20:59:23 sprawl cogitare.core.model[2443] INFO Total number of parameters: 21,840
2018-02-02 20:59:23 sprawl cogitare.core.model[2443] INFO Starting the training ...
2018-02-02 21:02:04 sprawl cogitare.core.model[2443] INFO Training finished

Stopping training after 10 tries. Best score 0.0909
Model restored from: /tmp/model.pt

To check the model loss and accuracy on the validation dataset:

def model_accuracy(output, data):
    _, indices = torch.max(output, 1)
    
    return accuracy(indices, data[1])

# evaluate the model loss and accuracy over the validation dataset
metrics = model.evaluate_with_metrics(data_validation, {'loss': model.metric_loss, 'accuracy': model_accuracy})

# the metrics is an dict mapping the metric name (loss or accuracy, in this sample) to a list of the accuracy output
# we have a measurement per batch. So, to have a value of the full dataset, we take the mean value:

metrics_mean = {'loss': 0, 'accuracy': 0}
for loss, acc in zip(metrics['loss'], metrics['accuracy']):
    metrics_mean['loss'] += loss
    metrics_mean['accuracy'] += acc.item()

qtd = len(metrics['loss'])

print('Loss: {}'.format(metrics_mean['loss'] / qtd))
print('Accuracy: {}'.format(metrics_mean['accuracy'] / qtd))
Loss: 0.10143917564566948
Accuracy: 0.9846252860411899

One of the advantages of Cogitare is the plug-and-play APIs, which let you add/remove functionalities easily. With this sample, we trained a model with training progress bar, error plotting, early stopping, grad clipping, and model evaluation easily.

4. Contribution

Cogitare is a work in progress project, and any contribution is welcome.

You can contribute testing and providing bug reports, proposing feature ideas, fixing bugs, pushing code, etcs.

  1. You want to propose a new Feature and implement it
    • post about your intended feature, and we shall discuss the design and implementation. Once we agree that the plan looks good, go ahead and implement it.
  2. You want to implement a feature or bug-fix for an outstanding issue
    • Look at the outstanding issues here: https://github.com/cogitare-ai/cogitare/issues
    • Pick an issue and comment on the task that you want to work on this feature
    • If you need more context on a particular issue, please ask and we shall provide.

Once you finish implementing a feature or bugfix, please send a Pull Request to https://github.com/cogitare-ai/cogitare

If you are not familiar with creating a Pull Request, here are some guides:

Comments
  • [Feature request] Plugin to watch the training on web (tensorboard integrated with cogitare plugins)

    [Feature request] Plugin to watch the training on web (tensorboard integrated with cogitare plugins)

    • plot training error\std
    • plot validation error\std
    • time remaining
    • button to stop the training process
    • button to save the model at the current step
    • button to pause the training
    • button to resume the training
    • plot model parameters statistics
    • save/load model execution log, to compare and analyze different executions [1]
    • plot execution graph
    • maybe something like named-scope from tensorflow [2]
    • x-axis: by value or by relative time [3]
    • plot smothing
    • display real-time execution machine/gpu stats
    • add Hyper-parameter option to modify its value from the web interface

    [1] screenshot from 2017-10-31 17-04-09

    [2] screenshot from 2017-10-31 17-13-13

    [3] screenshot from 2017-10-31 17-52-29

    enhancement hard 
    opened by aron-bordin 1
  • [Feature Request] Implement History plugin

    [Feature Request] Implement History plugin

    A plugin that records all (or a fraction, if given a filter) of variables during the training process.

    It watches all hooks, capture the variables, and then can be exported.

    • be compatible with the Cogitare Monitor, implementing a history viewer.
    enhancement medium 
    opened by aron-bordin 0
  • [Feature Request] Add map parameter to dataholders

    [Feature Request] Add map parameter to dataholders

    A callable parameter, that can act over the sample before generating the batch.

    It should allow easy-to-use preprocessing algorithms through a distributed interface (threads, processes, machines)

    Add on dataholder:

    • on_sample_loaded
    • on_batch_loaded

    Add on asyncloader:

    • on_batch_loaded (useful for loading batches to gpu before using)
    enhancement 
    opened by aron-bordin 0
  • before first release, profile everything to make mem/speed improvements

    before first release, profile everything to make mem/speed improvements

    Logs.

    18/09 - replaced python indices by numpy indices and python shuffle by numpy shuffle in dataholder. In a dataset with millions of samples, improved by ~15x.

    enhancement 
    opened by aron-bordin 0
  • [Feature Request] add utils.auto_optim

    [Feature Request] add utils.auto_optim

    add a simple function on utils, which receives the optimizer name, the model parameters, and its arguments. This function will create the optimizer and return it.

    (if testing multiples optimizers, it's not required to change the code to change an optimizer. you can, for example, use an argument named "optim" and just pass this argument to the function)

    enhancement help wanted easy 
    opened by aron-bordin 0
  • [Feature Request] Implement Interactive SIGINT Interrupt

    [Feature Request] Implement Interactive SIGINT Interrupt

    A plugin that listens SIGINT signal during training.

    When receiving the signal, gives some options to the interactive user:

    • save/load the model state
    • quit training
    • maybe something else
    enhancement help wanted easy 
    opened by aron-bordin 0
Releases(v0.1.0)
  • v0.1.0(Feb 3, 2018)

    The first release of Cogitare.

    Support:

    • Model

    • Sequential Model

    • DataHolder

    • Sequential DataHolder

    • DataSet

    • Sequential DataSet

    • AsyncDataLoader

    • Metrics (classification, spatial)

    • Classic Models (LR, MLP)

    • Web Monitor (system usage, system details)

    • Early stopping plugin

    • Evaluator plugin (different test metrics on the model)

    • Logger

    • Plotting (matplotlib)

    • Progress Bars

    • Some utilities

    • Documentation with examples

    • Tests: 92% of coverage (8% remaining is of the Monitor undefined interface)

    Source code(tar.gz)
    Source code(zip)
Owner
Cogitare - Modern and Easy Deep Learning with Python
A modern, fast, and modular deep learning and machine learning framework for Python
Cogitare - Modern and Easy Deep Learning with Python
Gated-Shape CNN for Semantic Segmentation (ICCV 2019)

GSCNN This is the official code for: Gated-SCNN: Gated Shape CNNs for Semantic Segmentation Towaki Takikawa, David Acuna, Varun Jampani, Sanja Fidler

859 Dec 26, 2022
A minimal TPU compatible Jax implementation of NeRF: Representing Scenes as Neural Radiance Fields for View Synthesis

NeRF Minimal Jax implementation of NeRF: Representing Scenes as Neural Radiance Fields for View Synthesis. Result of Tiny-NeRF RGB Depth

Soumik Rakshit 11 Jul 24, 2022
ML-Decoder: Scalable and Versatile Classification Head

ML-Decoder: Scalable and Versatile Classification Head Paper Official PyTorch Implementation Tal Ridnik, Gilad Sharir, Avi Ben-Cohen, Emanuel Ben-Baru

189 Jan 04, 2023
The official code for paper "R2D2: Recursive Transformer based on Differentiable Tree for Interpretable Hierarchical Language Modeling".

R2D2 This is the official code for paper titled "R2D2: Recursive Transformer based on Differentiable Tree for Interpretable Hierarchical Language Mode

Alipay 49 Dec 17, 2022
Physical Anomalous Trajectory or Motion (PHANTOM) Dataset

Physical Anomalous Trajectory or Motion (PHANTOM) Dataset Description This dataset contains the six different classes as described in our paper[]. The

0 Dec 16, 2021
ALFRED - A Benchmark for Interpreting Grounded Instructions for Everyday Tasks

ALFRED A Benchmark for Interpreting Grounded Instructions for Everyday Tasks Mohit Shridhar, Jesse Thomason, Daniel Gordon, Yonatan Bisk, Winson Han,

ALFRED 204 Dec 15, 2022
Official pytorch implementation of "Feature Stylization and Domain-aware Contrastive Loss for Domain Generalization" ACMMM 2021 (Oral)

Feature Stylization and Domain-aware Contrastive Loss for Domain Generalization This is an official implementation of "Feature Stylization and Domain-

22 Sep 22, 2022
Code for the ICCV2021 paper "Personalized Image Semantic Segmentation"

PSS: Personalized Image Semantic Segmentation Paper PSS: Personalized Image Semantic Segmentation Yu Zhang, Chang-Bin Zhang, Peng-Tao Jiang, Ming-Ming

张宇 15 Jul 09, 2022
Image-to-image translation with conditional adversarial nets

pix2pix Project | Arxiv | PyTorch Torch implementation for learning a mapping from input images to output images, for example: Image-to-Image Translat

Phillip Isola 9.3k Jan 08, 2023
Gradient Step Denoiser for convergent Plug-and-Play

Source code for the paper "Gradient Step Denoiser for convergent Plug-and-Play"

Samuel Hurault 11 Sep 17, 2022
For IBM Quantum Challenge Africa 2021, 9 September (07:00 UTC) - 20 September (23:00 UTC).

IBM Quantum Challenge Africa 2021 To ensure Africa is able to apply quantum computing to solve problems relevant to the continent, the IBM Research La

Qiskit Community 48 Dec 25, 2022
Code for the paper "Zero-shot Natural Language Video Localization" (ICCV2021, Oral).

Zero-shot Natural Language Video Localization (ZSNLVL) by Pseudo-Supervised Video Localization (PSVL) This repository is for Zero-shot Natural Languag

Computer Vision Lab. @ GIST 37 Dec 27, 2022
Voice Gender Recognition

In this project it was used some different Machine Learning models to identify the gender of a voice (Female or Male) based on some specific speech and voice attributes.

Anne Livia 1 Jan 27, 2022
Efficient Training of Visual Transformers with Small Datasets

Official codes for "Efficient Training of Visual Transformers with Small Datasets", NerIPS 2021.

Yahui Liu 112 Dec 25, 2022
Official pytorch code for SSAT: A Symmetric Semantic-Aware Transformer Network for Makeup Transfer and Removal

SSAT: A Symmetric Semantic-Aware Transformer Network for Makeup Transfer and Removal This is the official pytorch code for SSAT: A Symmetric Semantic-

ForeverPupil 57 Dec 13, 2022
The Generic Manipulation Driver Package - Implements a ROS Interface over the robotics toolbox for Python

Armer Driver Armer aims to provide an interface layer between the hardware drivers of a robotic arm giving the user control in several ways: Joint vel

QUT Centre for Robotics (QCR) 13 Nov 26, 2022
ViSD4SA, a Vietnamese Span Detection for Aspect-based sentiment analysis dataset

UIT-ViSD4SA PACLIC 35 General Introduction This repository contains the data of the paper: Span Detection for Vietnamese Aspect-Based Sentiment Analys

Nguyễn Thị Thanh Kim 5 Nov 13, 2022
TensorFlow (v2.7.0) benchmark results on an M1 Macbook Air 2020 laptop (macOS Monterey v12.1).

M1-tensorflow-benchmark TensorFlow (v2.7.0) benchmark results on an M1 Macbook Air 2020 laptop (macOS Monterey v12.1). I was initially testing if Tens

particle 2 Jan 05, 2022
Modifications of the official PyTorch implementation of StyleGAN3. Let's easily generate images and videos with StyleGAN2/2-ADA/3!

Alias-Free Generative Adversarial Networks (StyleGAN3) Official PyTorch implementation of the NeurIPS 2021 paper Alias-Free Generative Adversarial Net

Diego Porres 185 Dec 24, 2022
1st Place Solution to ECCV-TAO-2020: Detect and Represent Any Object for Tracking

Instead, two models for appearance modeling are included, together with the open-source BAGS model and the full set of code for inference. With this code, you can achieve around 79 Oct 08, 2022