Radal - A no-code platform for customizing language models.

RadalBeta v0 offers a no-code solution for training small language models (SLMs), making it accessible for users to create and deploy AI models without programming knowledge. The platform allows users to drag and drop their data, interact with an AI copilot, and iterate visually on their training processes. With features like Hugging Face integration and the ability to export quantized models, RadalBeta is designed for various applications across industries.

1. Interactive AI Layer

   Engage with the AI as it constructs a tailored training flow, simplifying the model training process.

2. No Code Canvas

   Users can edit and iterate on their models quickly without needing to write any code.

3. Hugging Face Integration

   Automatically push trained models to the Hugging Face hub for easy sharing and collaboration.

4. Local Inference

   Trained models can run on edge devices, allowing for continued inference without relying on cloud services.

5. Training Summary

   Access model settings, training statistics, and download the .gguf file instantly for further use.

1. Start by visiting the RadalBeta website and signing up for an account.
2. Prepare your dataset and drag it into the platform.
3. Use the AI copilot to guide you through the training setup.
4. Customize your model parameters as needed.
5. Once training is complete, export your model for deployment.

What size models can I train with Radal?

Radal allows users to train various sizes of small language models, depending on the dataset and application requirements.

Can I use my own base model?

Yes, users can utilize their own base models to fine-tune and adapt them for specific tasks.

What does the Copilot actually do?

The Copilot assists users in creating a tailored training flow, providing guidance and suggestions throughout the process.

How can I try out my trained model?

Trained models can be tested locally on edge devices or through the Radal platform for immediate feedback.

Why even build this? Don’t LLMs already do everything?

Radal focuses on small language models that can be fine-tuned for specific applications, offering advantages in efficiency and privacy compared to larger, general-purpose models.

Price data is not available yet; please visit the official Radal website for the latest pricing information.

1. RadalBeta provides a user-friendly interface that simplifies the process of training small language models, making it accessible for users without coding skills.
2. The integration with Hugging Face and the ability to run models locally are significant advantages for users concerned about data privacy and model deployment.
3. However, the platform may require users to have a basic understanding of AI concepts to fully leverage its capabilities.
4. While RadalBeta is innovative, it may not yet support the complexity needed for more advanced AI applications, which could limit its appeal to experienced developers.