The problem Milon : Federation Simplified solves

In the field of machine learning, training complex and resource-intensive models often requires substantial computational power and specialized hardware, such as high-end GPUs or large-scale computing clusters. This presents a significant challenge for individuals, research groups, or organizations with limited access to such resources. The inability to train state-of-the-art models due to hardware constraints can impede progress in various domains, including computer vision, natural language processing, and scientific research.

Moreover, traditional centralized model training approaches raise privacy concerns, as they typically involve aggregating sensitive data on a central server. This centralization of data not only poses potential risks of data breaches or unauthorized access but also conflicts with data protection regulations and ethical principles.

Challenges we ran into

Sending Epochs over the blockchain network , Getting the approx weights of the epoch , Working with Rust , Smart contract , Flag sending for when epoch when ready