The field of Artificial Intelligence (AI) is advancing at an exponential rate, and one of the fascinating areas is image generation. A Discord channel that specializes in generating images had an exciting conversation regarding their new updates. In this blog post, we will delve into the new updates, features, and improvements that they discussed.
The first update that the Discord channel discussed was the release of version 2. The team had been working hard to put together the model, and as users have noticed, V2 comes with an illustrated artistic style. However, some bugs have been discovered, and the team is working hard to fix them. In case you still want to use V1, you can easily do that by typing "Dash dash V one." Furthermore, the team is already working on their third model, which they hope to release in a couple of months. According to Floripa, the third model is expected to have significant improvements that users will be happy with.
Another exciting update that the Discord channel discussed was the upcoming release of their web gallery. The gallery will allow users to store all their images created on the platform easily. Users can view, download, and even share their creations with others. The web gallery will be released in a week, and the team is looking forward to getting feedback from users.
The Discord channel also revealed that they are working on implementing a ranking and point reward system. However, they did not provide much information about it but promised to give more details on both their Twitch channel and Discord.
One user asked if there were any plans to implement seed functionality. In response, BlueWillowAI explained that while the team plans to implement it, it won't be anytime soon. The reason being that the model has been iterating frequently, and keeping the seed functioning across all versions would be challenging. However, they still plan to work on it in the future.
Another user inquired if the AI would be able to understand and generate images from emojis. BlueWillowAI stated that while some emojis might work, they have not optimized the AI to use emojis as prompts. They promised to look into it and make sure that the feature is solid.
One user also asked about the base resolution of the AI. BlueWillowAI explained that the default resolution was 512 by 512. However, the AI's resolution was limited by GPU computing constraints and some model constraints. The team has implemented an upscale model that users can take advantage of to improve the resolution of their images.
The Discord channel specializing in image generation has exciting updates that users can take advantage of. The release of version 2 and the upcoming release of the web gallery are some of the significant improvements that users can expect. The team is also working on implementing a ranking and point reward system, while also promising to optimize the AI to use emojis as prompts. Overall, these updates promise to make the platform more user-friendly and fun to use.