Start building useful computer vision models in under 30 minutes
Epigos equips you with all the essentials to transform images into AI models. We’ve developed a comprehensive set of tools to make computer vision accessible, even for those without extensive machine learning expertise.
This guide will lead you through the process of training a computer vision model that can detect different types of blood cells in a blood histology image. The insights gained here can be applied to construct a customized computer vision model tailored to your specific needs.
To begin, initiate your journey by signing up with either your email or Google account. You’ll then be redirected to the dashboard, where you can create a fresh workspace under your account.
Next, click on Create new project
to start a new project that will contain the resources for your model training.
Provide a name, select the type of model, and describe the model you wish to build.
The labels
for your model, as well as annotation instructions for your data labelers, must then be provided.
After you’ve created your project, go to the project details by clicking on the project card.
Click the Uploads
link in the sidebar, then click Upload dataset
in the top right corner of the uploads
page.
A dialog box will appear, allowing you to upload files from your computer.
Drop or upload files into the dialog area, and click on Upload files
to upload the selected files.
When the upload is completed, the uploaded batch will be displayed in table.
To annotate, click on the Annotate
button to start annotating your dataset batch.
The annotation page will display the image to be annotated, together with the class labels to assign to the images.
To annotate the image follow these steps:
Training a model within Epigos AI is streamlined with a single click. We take care of GPU allocation and associated expenses, and additionally provide you with readily available optimized deployment choices, which will be elaborated on in subsequent sections of this guide.
Once you’ve created your dataset, the option to initiate model training becomes visible. To begin the training process, kindly follow the steps outlined below:
Train model
Start model training
. Your will begin training on our platform as background task. You’ll be notified once your model training is completed.After your model is trained, you can review the training metrics by clicking on the model link in the models
table.
Training a model can take from 1 to 12hrs depending on the size of your data. The process is done in the background and we’ll notify you once the model is ready to be used.
We offer a hosted inference UI that you can use to test your model on real data.
The model you’ve trained on Epigos AI is fully optimized and configured for use across various deployment choices. If you’re uncertain about the most suitable deployment path for your model, we offer a valuable guide to aid you in making an informed decision.
Test your model in the dashboard
Run predictions directly in the models details page to evaluate the evaluate the performance of your models on sample images.
Use the hosted web app
Generate a URL to the hosted web UI to make predictions on your model.
Anyone with access to the link can run predictions on your model in the web app.
Use API or SDKs
Use the Epigos REST API, Python and Node.js SDK to manage and use your models.
Start building useful computer vision models in under 30 minutes
Epigos equips you with all the essentials to transform images into AI models. We’ve developed a comprehensive set of tools to make computer vision accessible, even for those without extensive machine learning expertise.
This guide will lead you through the process of training a computer vision model that can detect different types of blood cells in a blood histology image. The insights gained here can be applied to construct a customized computer vision model tailored to your specific needs.
To begin, initiate your journey by signing up with either your email or Google account. You’ll then be redirected to the dashboard, where you can create a fresh workspace under your account.
Next, click on Create new project
to start a new project that will contain the resources for your model training.
Provide a name, select the type of model, and describe the model you wish to build.
The labels
for your model, as well as annotation instructions for your data labelers, must then be provided.
After you’ve created your project, go to the project details by clicking on the project card.
Click the Uploads
link in the sidebar, then click Upload dataset
in the top right corner of the uploads
page.
A dialog box will appear, allowing you to upload files from your computer.
Drop or upload files into the dialog area, and click on Upload files
to upload the selected files.
When the upload is completed, the uploaded batch will be displayed in table.
To annotate, click on the Annotate
button to start annotating your dataset batch.
The annotation page will display the image to be annotated, together with the class labels to assign to the images.
To annotate the image follow these steps:
Training a model within Epigos AI is streamlined with a single click. We take care of GPU allocation and associated expenses, and additionally provide you with readily available optimized deployment choices, which will be elaborated on in subsequent sections of this guide.
Once you’ve created your dataset, the option to initiate model training becomes visible. To begin the training process, kindly follow the steps outlined below:
Train model
Start model training
. Your will begin training on our platform as background task. You’ll be notified once your model training is completed.After your model is trained, you can review the training metrics by clicking on the model link in the models
table.
Training a model can take from 1 to 12hrs depending on the size of your data. The process is done in the background and we’ll notify you once the model is ready to be used.
We offer a hosted inference UI that you can use to test your model on real data.
The model you’ve trained on Epigos AI is fully optimized and configured for use across various deployment choices. If you’re uncertain about the most suitable deployment path for your model, we offer a valuable guide to aid you in making an informed decision.
Test your model in the dashboard
Run predictions directly in the models details page to evaluate the evaluate the performance of your models on sample images.
Use the hosted web app
Generate a URL to the hosted web UI to make predictions on your model.
Anyone with access to the link can run predictions on your model in the web app.
Use API or SDKs
Use the Epigos REST API, Python and Node.js SDK to manage and use your models.