GPT3Discord VS cog-stable-diffusion

git clone https://github.com/Kav-K/GPT3Discord.git

Hey everybody! I posted on this subreddit earlier about a project of mine, https://github.com/Kav-K/GPT3Discord, a full fledged Discord-OpenAI integration, since then, it has gained a lot of popularity!

I've used https://github.com/Kav-K/GPT3Discord with custom indexes functionality. It's pretty cool.

Kav-K/GPT3Discord: A robust, all-in-one GPT3 interface for Discord. ChatGPT-style conversations, image generation, AI-moderation, custom indexes/knowledgebase, youtube summarizer, and more! (github.com)

Hey all! I'm sure those who frequent this sub have seen my posts before, I'm posting again about my project GPT3Discord (https://github.com/Kav-K/GPT3Discord), a fully fledged OpenAI interface for discord that provides infinite-context chatting with GPT3 with permanent memory, image generation, ai-assisted google search, document indexing, AI-based moderation, translations, language detection, and much more.

Look at AI Dungeon or KoboldAI, there are some tricks there as others mentioned. Also look here, this one seems to have some cool ideas as well: https://github.com/Kav-K/GPT3Discord

Stable Diffusion

stable-diffusion: A latent text-to-image diffusion model capable of generating photorealistic images given any text input

Next, make a Flask app so your app can receive the inbound text message to your Twilio phone number. That string is then passed to Replicate 's stable diffusion model to generate photo-realistic images given that prompt.

The Azure Function App executes a serverless function that: a) Runs the text summary model (Azure Cognitive Service for Language) b) Runs the image generation model (stable-diffusion on Replicate AI) c) Uploads the image to Azure Blob Storage

I'm testing prompts online because it's faster than my computer: https://replicate.com/stability-ai/stable-diffusion

There isn't a direct integration with Zapier, but you can use the Webhooks & API calls within Zapier to connect to Stable diffusion. The simplest way is to do it through Replicate: https://replicate.com/stability-ai/stable-diffusion

I ended up using stable diffusion from replicate. They don't have a filter on the input text, they just analyse the output for NSWF and block it if it seems to be NSFW. So sometimes NSWF-ish pictures may slip through.

I'm getting same-ish results as mine with replicate. They don't have PLMS as an option and euler trips that silly porn filter on the dressed pirates for whatever reason, but it seems to closely match.