본문 바로가기

Create better performing AI with more data with Jonathan®'s federated learning technology

Create model training

Analyze patient emotions, predict depression, and monitor treatment progress based on 34 types of emotion classification models and natural language processing technology
Joint development with the Department of Psychiatry at Catholic University of Seoul St. Mary’s Hospital

Hyperparameter search

In HPS tasks, target hyperparameters are searched for. A hyperparameter that yields the best score is picked and the optimal value is derived.

Hyperparameter search

As HPS operates, the learning performance of the model for each round is improved and a visualization function is provided.

Check logs for each training

You can check logs and graphs for the learning results for each task.

Deployment worker management

When entering the details page of the deployment project, you can add/remove new workers or check the status of the workers.
When creating multiple workers, load-balancer is automatically applied to provide smooth service.

Deployment worker dashboard

This is the dashboard of the deployment project. You can check the status and records of the distribution system and API.

Test your deployment model

On the test details page, you can input data into the deployed model and check the results inferred by the model as output.
In the same way as the testing method, it can be utilized by sending an API request from an application that requires AI utilization.

Provides LLM testing and LLMOps features

LLM models developed through the platform can be tested
Supports fine-tuning and RAG functions to customize the LLM model to suit the client's requirements