Sagify

Sagify is a command-line utility designed to simplify and streamline the training and deployment of machine learning and deep learning models on AWS SageMaker. It caters to developers of all levels, from beginners to experts, and eliminates the need for a team dedicated to implementing ML tools for data scientists, streamlining the process and increasing collaboration among users. Sagify automates the process of creating necessary file structures, making it easier for the user to concentrate on coding, and offers support for Bayesian Hyperparameter Optimization, allowing users to tune their models’ hyperparameters automatically. Sagify serves as an end-to-end solution for machine learning model deployment, training, and monitoring by seamlessly integrating with Aporia or Superwise. Compatible with AWS SageMaker, Sagify provides users with flexibility in pricing plans, with a freemium tier available that allows monitoring of up to three models. Sagify remains a valuable tool for developers through its continuously improving codebase and user-friendly interface.

TLDR

Sagify is a tool that simplifies the training and deployment of ML and deep learning models on AWS SageMaker. With a command-line interface, users can complete the entire process, including customizing templates for multiple programming languages, within a day. The tool caters to users of all levels and automatically creates file structures, outlining clear project plans, and supports Bayesian Hyperparameter Optimization for automating the tuning of hyperparameters. Additionally, Sagify integrates with Aporia or Superwise, providing users with an end-to-end solution for deploying, training, and monitoring models. Compatible with AWS SageMaker and maintaining a flexible pricing plan, Sagify is a valuable tool for developers of all levels.

Company Overview

Sagify is a command-line utility that simplifies and facilitates the training and deployment of machine learning and deep learning models on AWS SageMaker. This tool allows users to complete the entire process, from training to deployment, in a matter of hours. Sagify eliminates the need for a team dedicated to implementing ML tools for data scientists, streamlining the process and facilitating collaboration. Sagify has an open-source code that is compatible with a range of programming languages, making it a practical tool for developers.

Sagify allows users to train, tune, and deploy machine learning models with ease, eliminating the need for a wealth of technical knowledge. Users can complete the entire process within a day, using a relatively simple command-line interface. Additionally, Sagify offers support for Bayesian Hyperparameter Optimization, allowing users to automate the tuning of their model's hyperparameters, resulting in quicker and more accurate models.

Sagify provides a clear and concise structure, allowing users to focus on the necessary actions as data scientists, without worrying about the rest. The tool automatically creates the necessary file structures and outlines a clear project plan, allowing users to focus on the most critical part of their work - coding. Additionally, the integration of Sagify with Aporia or Superwise provides users with an end-to-end solution for model monitoring, allowing developers to quickly and easily deploy, train and monitor their models.

Sagify is compatible with AWS SageMaker, allowing users to take full advantage of the platform's flexibility and scalability. Sagify also maintains a flexible pricing plan, with a freemium tier available to users, allowing them to monitor up to three models. Sagify's continuously improving codebase and user-friendly interface make it a valuable tool for developers of all levels, be it novices or experts.

Features

Simplified Training and Deployment Process

Easy-to-Use Command-Line Interface

Sagify offers a simple command-line interface that enables users to train, tune, and deploy machine learning models quickly and efficiently. The interface is easy to navigate, simplifying the overall training and deployment process for users who may not have a wealth of technical knowledge. With Sagify, even novice developers can complete the entire process within a day, including training and deploying their models.

Automated File Structures and Project Plans

Another benefit of Sagify is its ability to automatically create file structures and project plans necessary for training and deploying machine learning models. This feature simplifies the overall process, allowing users to focus on the most critical part of their work - coding. As a result, Sagify provides a clear and concise structure and allows users to complete necessary actions as data scientists, without worrying about the rest.

Integration with Model Monitoring Tools

Sagify offers seamless integration with Aporia or Superwise, providing users with an end-to-end solution for model monitoring. As a result, developers can quickly and easily deploy, train and monitor their models, ensuring that the models perform optimally throughout the process.

Hyperparameter Optimization and Customizable Templates

Automated Hyperparameter Optimization

Sagify's Bayesian Hyperparameter Optimization feature allows users to automate the tuning of their model's hyperparameters, resulting in quicker and more accurate models. This feature comes in handy for data scientists who may not have a wealth of technical knowledge in this area, allowing them to streamline the process by automating it.

Customizable Templates for Popular ML Frameworks

Sagify offers customizable templates for a range of programming languages, including Python, R, and Tensorflow, which make it easy for developers to start with popular ML frameworks, reducing the learning curve.

Open-Source Codebase

Sagify has an open-source code that is compatible with a range of programming languages, making it a practical tool for developers at all levels; from novice to expert. This feature provides flexibility, allowing developers to adapt the tool to meet their specific needs.

Flexible Pricing Plans and Compatibility with AWS SageMaker

Freemium Tier with Model Monitoring for up to Three Models

Sagify maintains a flexible pricing plan, with a freemium tier available to users. This tier allows users to monitor up to three models, enabling them to experiment with the tool before upgrading to a higher level plan.

Compatible with AWS SageMaker

Sagify is compatible with AWS SageMaker, allowing users to take full advantage of the platform's flexibility and scalability. As a result, users can complete the entire process, from training to deployment, in a matter of hours, without worrying about having to use different tools and platforms.

Continuously Improving Codebase

Sagify's continuously improving codebase allows for regular updates and improvements, ensuring that users have access to the latest features and functionalities. This feature ensures that Sagify remains a valuable tool for developers of all levels, whether they are novices or experts.

Increased Efficiency and Time-Saving Benefits

Fully Automated ML Pipeline

Sagify automates the entire ML pipeline process, eliminating the need for a team dedicated to implementing ML tools for data scientists, streamlining the process, and facilitating collaboration. As a result, Sagify saves developers time and increases their efficiency.

Quicker Model Training and Deployment

With Sagify, users can complete the entire process, from training to deployment, in a matter of hours. This feature makes it an attractive tool for users who need to train and deploy models quickly without having to worry about the technicalities of it.

More Accurate Models

By automating the tuning of models' hyperparameters with Sagify's Bayesian Hyperparameter Optimization feature, users can train more accurate models. This feature saves developers time while ensuring that the models perform optimally throughout the process.

Integrations

Sagify provides users with the ability to utilize the Superwise platform to define workflows that automatically monitor different AI use cases such as data drift, performance degradation, data integrity, and model activity. Sagify users can monitor up to three models using the free tier of Superwise. To set up Superwise, users should create an account in the platform by clicking on the Account button. Once the account is created, users can access the Personal tokens option in the User profile to create an access token on the Superwise dashboard.

Using Superwise with Sagify

Users can use the Superwise SDK to create the model, and then initialize Sagify using a command: sagify init. The command will prompt users to enter the SageMaker app name, which should be iris-model, and whether they are starting a new project, to which users should answer y. Next, users should make sure to choose Python version 3 and the AWS profile and region they want to use. The requirements.txt path should be entered as an answer to the prompt “Type in the path to requirements.txt.”

In the src directory, Sagify creates a module called “sagify_base” which needs to have requirements.txt with the following content:

numpy==1.19.2
            pandas==1.1.3
            scikit-learn==0.24.2

The Iris data set should also be downloaded and saved under src/sagify_base/local_test/test_dir/input/data/training/ in a file named “iris.data”. Users can then replace the “TODOs” in the train(...) function with the following text:

file_size_limitation = 30000 # in rows
            filepath = os.path.join(TRAINING_DIR, "iris.data")
            df = pd.read_csv(filepath, header=None, names=["sepal_length", "sepal_width", "petal_length", "petal_width", "class"])
            if len(df) > file_size_limitation:
                df = df.sample(file_size_limitation)
            X_train, _, y_train, _ = split_data(df, 0.3)

Add the following code to the top of the file src/sagify_base/training/training.py:

import os
            from sklearn.model_selection import train_test_split as split_data
            import pandas as pd

Within the src/sagify_base/prediction/prediction.py file, the body of the predict(...) function should be replaced with the following:

input_data = json.loads(input_data)
            instances = input_data['instances']
            preds = model.predict(instances)
            return preds.tolist()

In the same file, within the ModelService class, the body of the get_model() function should be replaced:

model_path = os.path.join(model_dir, 'model.joblib')
            model = joblib.load(model_path)

A new function should also be added to the ModelService class using the following code:

def get_superwise_model(self):
                return self.get_model()

The file should have the following line added to the top:

import joblib
            import os
            import json

After users have modified the necessary files, they can build and train the ML model using the command sagify build and then sagify local train. Once the model is deployed following the command sagify local deploy, users can call the inference endpoint using a curl command. With these steps completed, Sagify users can monitor data and model activity on Superwise dashboards.

FAQ

What is Sagify?

Sagify is a command-line utility that simplifies and streamlines the training and deployment of machine learning and deep learning models on AWS SageMaker. It allows users to complete the entire process, from training to deployment, in a matter of hours.

Who is Sagify for?

Sagify is designed to be user-friendly and accessible to developers of all levels, from novices to experts. It is a practical tool for developers who need to train, tune, and deploy machine learning models with ease, without requiring a wealth of technical knowledge. Additionally, Sagify provides a solution for data scientists who need to collaborate by facilitating model deployment and monitoring.

What programming languages are compatible with Sagify?

Sagify is an open-source code compatible with a range of programming languages, including Python, Java, Node.js, Ruby, R, and Go, among others. This compatibility makes Sagify a valuable tool for developers who use multiple languages.

How does Sagify simplify the machine learning process?

Sagify provides a clear and concise structure that allows users to focus on the necessary actions as data scientists without worrying about the rest. The tool automatically creates the necessary file structures and outlines a clear project plan, allowing users to focus on the most critical part of their work - coding. Additionally, Sagify offers support for Bayesian Hyperparameter Optimization, which automates the tuning of model's hyperparameters, resulting in quicker and more accurate models.

What is Bayesian Hyperparameter Optimization?

Bayesian Hyperparameter Optimization is a technique used to tune the hyperparameters of machine learning models automatically. It eliminates the need for manual hyperparameter tuning, and it can be done quickly and easily by a data scientist using Sagify.

What are the benefits of using Sagify?

Sagify has several benefits that make it a valuable tool for developers and data scientists. First, Sagify simplifies the machine learning process and eliminates the need for extensive technical knowledge. Second, Sagify is an open-source code that is compatible with a range of programming languages. Third, Sagify offers support for Bayesian Hyperparameter Optimization, allowing users to automate the tuning of their model's hyperparameters. Fourth, Sagify integrates with Aporia or Superwise for model monitoring, providing an end-to-end solution for deploying, training and monitoring models.

What is the pricing plan for Sagify?

Sagify maintains a flexible pricing plan, with a freemium tier available to users, allowing them to monitor up to three models. The tool is reasonably priced and is an excellent option for developers looking for an affordable solution for model deployment and monitoring on AWS SageMaker.

Is Sagify compatible with AWS SageMaker?

Yes, Sagify is compatible with AWS SageMaker, allowing users to take full advantage of the platform's flexibility and scalability. Sagify simplifies the machine learning process on AWS, making it accessible to developers with varying levels of technical expertise.

How can I get started with Sagify?

To get started with Sagify, simply download the tool from the website and install it. Sagify provides a user-friendly interface, and it is compatible with a range of programming languages. After installation, data scientists can begin by creating a new project and following the steps outlined by Sagify.