The problem NFTree solves

NFTree is essentially aims to bridge the gap between the extensive research done in the Image Processing sector of AI with Web3 through the means of NFTs. It eases the process of creating and customizing their NFTs using powerful Deep Learning models of OpenAI, Stable Diffusion, etc. Additionally, NFTree introduces an ERC721-based NFT called (of course!) NFTree (NFR). The NFR token is especially unique as it can be not only be minted using Generative AI models, but also can be combined with one another as well as split up into variations all while retaining the history of how the token was created !!

Challenges I ran into

While conceptualizing the building of NFTree and how it should work was relatively simple, the actual implementation part was far from it. Writing the smart contract to save the history of the token and how it was built to this point had me researching on memory management in Solidity and time-gas tradeoffs. Deciding the NFT standard to ahead with also took a while. Additionally, building my own models for the various tasks such as merging images and generating variations, as well as integrating the various models alreadya vailable proved to a rather time-consuming, but extremely informative and fun tasks