Easy AI Influencer Content with GPT-2 Image!
Opening up to paid members first, then free members soon!
I've heard a lot of good stuff about the GPT-2 Image model and, I had a rough start at first, but I was able to figure out how to make the most of it for AI influencer style content! To help, I've made a Free GPT available to anyone with an OpenAI account to use! No long prompts needed!
Silent Snow AI Influencer GPT-2 Image Generator: https://chatgpt.com/g/g-69ed5ffa1498819195c02895aaa14749-silent-snow-gpt-2-ai-influencer-image-generator
Just follow these few easy steps:
Upload at least two reference images of your character, one should be a headshot, and another should be more body focused. Specify any additional details you want to maintain consistency in (for example, a specific smartphone type and color).
GPT will confirm it's understanding and analysis of the reference images you provide. Add your prompt and let it ride!
Here are some of the images I was able to generate!
Ground Truth Character Images
P.S. I show you how to generate images just like this and make your AI influencer from scratch in the course we're making ;)
GPT-2 Generated Images
GPT-2 Image VS Nano Banana
Right now, I don't think I've tested it enough to make a decision, but so far I'm liking the results from GPT image 2
At the end of the day, I think image models are starting to plateau. We're starting to see the best of the best come to light, and the competitive edge is no longer in just about making a better, more capable model, it's about pricing and accessibility.
Right now, aside from generation quality, GPT Image 2 is winning for me purely due to the fact that it's extremely easily accessible, fairly quick, and relatively cheap compared to Nano Banana.
Even looking at costing of GPT Image 2 inside of ComfyUI, sure the high quality generation is quite expensive relative to NB2, but we have upscalers that can easily get you the scale that you need.
A Note on AI Influencer Course
I've been grinding a way on it which is why there haven't been too many videos out lately, but this one was just too quick and high quality to pass up!
The goal is to have the first SFW AI Influencer course finished by the end of the first week of May!
If there's any single piece of advice I could give right now, image quality will not be the problem for AI Influencers moving forward, strategy and positioning will be.
Comments (2)
Will the influencer course focus soley on paid models for images and video or will you also be showing how to do it with open source models like Z Image, Flux and LTX 2.3.
Great questions! There are technically three different courses that I'm working on. The main one is primarily SFW and getting started with open-source and closed-source models like Flux Klein, GPT Image 2, Nano Banana, Kling, LTX, and Seedance. The second course is LoRa training which focuses on creating your dataset and training a model on Wan, Klein, and ZIT. Finally, the third course will be NSFW focused and is largely a combination of the previous two courses combined. Without the knowledge from both of those, NSFW can be quite tricky sometimes. As of now, I'm almost finished with just the first part of the course for SFW (Klein, Nano, etc.), but have not started the other two. LTX is also not included in the base course as of now, but I'm working on it, been having problems with every LTX workflow I find for some reason regardless of where I'm running Comfy, it's been quite bizarre. The main goal of this first course that I'm working on is to be an overall comprehensive view on AI influencers, how the AI space works, step by step how I created my character with the workflows and tools in the course, and what you need to know beyond just generating high quality images & video (how to think about your character creation, how to make your accounts to prevent getting banned, etc.). The base course is also planned to have a custom node pack and some other tools specific to people in the course that won't be featured here! Hope this helps!