Picsellia

Welcome to Picsell.ia, your new go-to platform for your AI projects. It doesn't matter if you are an ML professional, an AI researcher or an aspiring data scientist, we got you covered.

What is Picsell.ia ?

Our mission is to give you all the necessary tools to relief the burden of AI projects off of your shoulders. As a data scientist / ML engineer / Researcher, you shouldn't have to worry about the following topics :

• Share your work with others

• Data annotation

• Data storage

• Infrastructure management

• Experiment tracking

• Model deployment

• Creating a knowledge base for your company

Fortunately, on Picsell.ia, we either give you the right tools to achieve the tasks above or even automate them if we can.

That's not all ! Because even if it's part of your job, we also decided to help you with the following points :

• State of the Art research

• Data versioning

• Visualize and analyze your Data

• Storing and versioning your models

Philosophy

As we know that most of you are using Jupyter Notebooks or tools like this to create your algorithms and share your work with your peers, Picsell.ia is fully integrated with your code thanks to a Python SDK that we developed.

But to allow you (and also non-tech people) to manage your projects easier, you can also use our our UI (the platform in fact) to perform the same actions as you would do with code.

From uploading data to remote experiment launch through project configuration, you can do everything either from the platform, or from your code.

We really hope you will find everything you need for your projects, if you have any feature request, please write us on one of the following channels. (You can talk with us in English or French, we prefer English though so everybody on the channels can understand 😌)

Contact

Mail : [e-mail]

Gitter : https://gitter.im/picsellia-team/community

Slack : https://join.slack.com/t/picsellbetaplatform/shared_invite/zt-nzostjdg-P4c3Li0TFN8T8lYcqBR6NQ (The link might expire so feel free to send us a mail for a new invite link)

Features

To help you figure out what you can or can't do with Picsell.ia, here are the main features that we developed that can cover many aspects of your projects, but don't hesitate to dig into each feature to see how you can achieve what you want or just ask us !

Data storage and versioning (Images only for now)

Isn't it a pain to find a way to store all your data and share it with your team in a way that it is accessible easily for all your experiments and projects ?

That's why we created the Datalake, where you can :

• Store your images

• Add tags

• Filter and visualize

• Select assets to create Datasets

To learn more about the Datalake :

To let you iterate over your data while being sure you don't lose everything in the way, we created Datasets which is a powerful versioning tool where you can :

• Create different version of datasets

• Annotate your data

• Clone images / labels / annotations from one dataset to another in a granular way

• Merge your annotations in one click

• Use model-assisted annotation

To learn more about Datasets :

Create a Dataset

If you want to work with other kinds of data such as text, sound, or tabular data, you can use the Experiment file-system (see the section below)

Data Annotation

If you stored some data in your Datalake and created one or several Datasets, you can now annotate your images with geometric shapes using our optimized tool or perform some model-assisted labeling using trained models.

To see all the possible way to annotate your data, please refer to this piece of documentation :

Create, annotate, review a Dataset

Experiment tracking

Have you ever feel that you are a notebook collector ? That they lack interactivity or on the contrary that you have no way to collect your logs, results efficiently for your algorithms when they run on remote machines ?

Well that's why we have designed a full-fledged experiment tracking system that allows you to perform a lot of things :

• Track your logs / metrics / results with one line of code

Log your results to Picsell.ia

• Store your files on Picsell.ia and retrieve them when or wherever you need them (again in one line of code)

Store your files to Picsell.ia

• Run any Python script (for example to train a model) on remote machines equipped with GPUs

• Iterate from previous experiments or from pre-trained models so you never have to configure anything

Initialize an experiment

Model HUB (or model ZOO)

As you might have noticed, our platform is community based, which means that you and every user can contribute to some pieces of Picsell.ia.

Our model HUB is one of those pieces.

The model HUB is a place where trained AI models can be stored (with all the files needed for configuration, training or inference) and that you can use as you please to do the following things :

• Use them to kickstart your experiments (for example fine-tuning a model on your own data)

• Deploy them for inference with an API endpoint

• Add some documentation so everyone knows what your model does and how to use it

• Share them with your team or organization

Models can be of any type or framework. You can save some Text to Speech Pytorch models as you can save Object detection Tensorflow models, it has no importance at all !

The cool thing is that if you are a fellow supporter of Open Source software, you can add your model to our Public HUB so every Picsell.ia user can use it on their own, or you can share it privately in your Organization HUB so only members of your team can access it for their projects.

Deploy models (Tensorflow only for now)

Now that you have trained your own custom model or using one from the model HUB, you might want to expose it to the world (or at least your company or client) with an API endpoint, hopefully you can do it with one click only from the Picsell.ia interface, or from the Python SDK and then make predictions using your API token.

If you want to learn more about deployment or how to deploy your models, you can go to this page :

Deploy model in production (Tensorflow only)

Use-cases

Here is a set of use-cases that describes some of the things you can achieve with Picsell.ia that might look like what you are trying to do in your project, if you think some use-cases are missing or are missing explanations, feel free to tell us !

Kick off your project 🚀

Start using Picsellia

Start with your use case 👨‍🏫

Create, annotate, review a DatasetCreate a new Dataset Version with merged labelsTrain a custom Object Detection modelDeploy model in production (Tensorflow only)

Learn more about Picsellia SDK 📚

Python SDK Reference