re:tune - the missing frontend for GPT-3

[median axle]

Hello OpenAI Community 👋

I am DJ, founder of re:tune and I am super excited to share https://retune.so

Get the most out of GPT-3 with our rich playground, create chat-bots, generate custom models from your data, and integrate in your workflow.

With re:tune, you can,

🤖 Turn any prompt into a chatbot.
👩‍💻 Get API to easily integrate into your app.
💁‍♀️ Pre-built templates for you to play around with.
🍳 Create custom fine-tuned models from your large dataset.
👩‍🍳 Modify your data and tweak parameters to get personalized output.
💅 Write a prompt once, and generate variable responses with our variables editor.

Coming soon 🤰
💰 Publish and monetize your apps
🍄 Feed your own data with embeddings
🦾 Supercharge your chatbots with internet access
🎼 Compose multiple prompts to build complex workflow

We have a discord server for faster support: https://discord.gg/WwzAfXVeNd

We are also live on product hunt: https://www.producthunt.com/posts/re-tune

[cerulean fable]

hey, the link to retune doesnt work

[mental umbra]

Genuine, monetisable usecase packed into a neat ui 👍 inspiring

[chrome herald]

I just had a try with https://retune.so/ Great product!

[drowsy lintel]

this actually seems pretty interesting ngl

[tawny kayak]

@median axle What's your recommendations for datasets to create chatbots for companies?

[signal parrot]

hi , what is the limitation of samples?

[feral idol]

@median axle mate, great job, honestly. keep it up

[flint tulip]

What does the Models tab do

@median axle

[median axle]

Thank you so much @feral idol

[median axle]

Thank you 🙌🙌

[median axle]

Heyo! What kind of chatbot are you looking to make?

[median axle]

There's no limitations during the beta

[median axle]

This tab is to create and list all the trained fine-tuned models. For now, we are manually activating fine-tuning features to make use only users who have large number of high quality data tries out that feature, because otherwise they might train with wrong expectations and end up getting poor results without realizing why.

If you want to train fine-tuned models, please drop me an email to [e-mail] with the account you signed up with. You can train 10 models with the free plan.

[flint tulip]

Nice. Yeah it was just an empty page ha

[drowsy lintel]

does this service require me to use an open ai api asewll ?

how does that end of it work? i am using a discord bot.. would love to train better personalities to it.. where are the api models held? how much will i have to modify my api inside of my bots programing?

[tawny kayak]

customer support Q&A kind of thing, maybe one for help with sales (product info, similar products etc)

[rigid galleon]

Very cool! With re:tune, can I fine-tune a model based on an export from a gmail inbox? How would that work?