file upload

Upload training file

SlyceAI accepts text based based training data containing input columns and the required output column.

No need to remove columns before import as all columns can be uploaded and only the columns required for training will be used.

As with all machine learning the training data needs to be as balanced as possible and clean of any duplicates before uploading.

Key features and benefits
  • CSV, Tab and Pipe delimited text files
  • Upload multiple files representing different models
  • Data parsed and checked
  • Column datatypes evaluated

Setup and Train

Choose either a Classification or Number Prediction model type.

Select one or more columns to use as inputs and a single column for the output.

Instruct SlyceAI to train the model, sit back and we will do the rest.

Once trained you can determine if the accuracy is acceptable, if not then you can adjust the combination of input columns and re-train without having to import the training file again.

Key features and benefits

  • Classification or Number Prediction models
  • Model accuracy calculated
  • Category columns for training efficiency
  • Quick and easy model setup
Categories
Prediction

Predict and Publish

Once trained the model can be executed in the UI using the built in SlyceAI prediction tool where you can manually test the model.

Publish the model to its own dedicated RESTful API endpoint (restricted with API keys) which can then be integrated into your software or website.

Key features and benefits

  • UI based model predictor for model testing
  • Full API swagger support
  • Integrate into your software as a RESTful API endpoint
  • API secured with keys

Sign up to the waiting list

SlyceAI is currently in pre-production - we would love to hear what you would like to build with SlyceAI and any features you'd like to see.

If you want to be involved or try out the pre-production version please fill out the contact form below.

Contact Us

We have received your message and will get back to you as soon as possible.

Error sending please try again