Radal is designed to provide enterprise users with a programming-free AI solution, with its core strengths being a drag-and-drop visualization interface and a system of preset templates. The platform supports the whole process from model architecture adjustment (e.g. adding attention mechanism) to training parameter setting, and users can view performance indicators such as accuracy rate and loss value in real time. The completed model can be exported to ONNX or TensorFlow format, which saves a lot of time and technical threshold compared with the traditional development mode.
This answer comes from the articleRadal: a low-code platform for rapid fine-tuning and optimization of AI modelsThe