🚀 Fine-Tuner.ai - Create Custom AI Models

Hello, fellow Bubblers! :star2:

:loudspeaker: We are excited to announce the launch of:

:one: Fine-Tuner.ai, an easy-to-use AI optimization platform that allows you to fine-tune your machine learning models without writing a single line of code :robot:
:two: Fine-Tuner.ai plugin for Bubble :electric_plug:

Fine-Tuner.ai provides a simple drag-and-drop interface to prepare your training data, choose a base model to fine-tune, set model hyperparameters, and train optimized AI models tailored to your needs.

And with the plugin you can:

  • Call your fine-tuned models directly from your Bubble workflows.
  • Submit new data samples to the existing training dataset directly from Bubble.
  • Fetch the training data associated with the current fine-tuned model.

Use Cases

  • Customer Acquisition: Improve targeting and marketing campaign effectiveness. Increase conversion rates and growth.
  • Predictive Analytics: Improve predictive analytics. Make data-driven decisions and gain business insights.
  • Personalization: Personalize customer experiences. Provide a seamless and engaging experience. Improve satisfaction.
  • Automation: Automate routine tasks. Free up time and resources to focus on other important tasks.
  • Financial Planning and Forecasting: Improve planning and forecasting. Make better decisions about investments, resources and future growth.
  • Customer Service: Provide fast and accurate customer service. Improve satisfaction and reduce support team workload.

Examples:

  1. Customer churn prediction: Identify potential churn and implement targeted retention strategies.
  2. Product recommendation: Provide personalized recommendations to enhance UX and increase sales.
  3. Image classification: Automatically categorize images based on content. Enable improved organization and search.
  4. Sentiment analysis: Analyze text to determine sentiment. Improve customer service and understand user feedback.
  5. Anomaly detection: Detect unusual patterns in data. Identify potential issues or opportunities.

Features:

  • No-code AI model creation: Easily build and optimize models without coding
  • Pre-trained AI models: Choose from pre-trained models for computer vision, NLP, predictions, recommendations and more (coming soon)
  • Optimized for your needs: Fine-tune models by training them on your data to tailor them to your use case
  • Simple drag-and-drop interface: Intuitive workflow to prepare data, select a model, set hyperparameters and train an optimized AI model
  • Integrate models into Bubble: Call your trained models directly from Bubble using the plugin
  • Community support: Get help from the Stack3 community if you have questions on using the plugin or training AI models
  • Always improving: We continuously release updates to the plugin and our AI models to provide the latest capabilities and best performance
  • Free to use: The plugin is freely available to all Bubble users to empower them with AI at no cost



:electric_plug:Check out the plugin here:

:books:API docs:
https://fine-tuner.ai/api_ref

:globe_with_meridians:Stack3 community:

Fine-Tuner.ai makes AI more accessible than ever before. Sign up for free at fine-tuner.ai to get started optimizing your AI today!

Let us know if you have any questions or need help getting started. We’d be happy to assist you!

4 Likes

I added a question ont he discord channel but thought I would post here in the forum as it may help those who dont want to join over there.

I have setup the plugin to test it out and have paid the subscription for the 'Hoby" plan that gives access to ta few more bit I wanted to test.

Installed the plugin, but also tested the API via the Bubble API connector per the API docs.

There is an error returned when I use the API connector, initialising and actually trying to use (as I was getting no errors back from the plugin when submitting).

Ive got my fine-tuner.ai API key plu the model ID which I have added when making may calls.

I created a new Text embedding endpoint:

Here is what I am getting back:

{“status”:“NOT_RUN”,“message”:“The condition for the workflow embeddings_upsert is not met. Workflow won’t run”}

Based on the message i an not setting something correctly… I just dont know what that is :slight_smile:
please help!

Hey, this should be fixed now. Please let me know if you still have any issues with this endpoint!

seems to have done the trick, thanks

1 Like

Our CSV Agent can now Visualize Data :bar_chart::exploding_head::rocket:

For example it can visualize data trends, financial forecasts, distribution of a variable, correlations between variables & more :fire:

Not just a query tool - it’s a powerful agent that can:

:one: interact with tools like the pandas data library :chart_with_upwards_trend:
:two: send queries to GPT4 to understand what to do :robot:
:three: can be used on larger databases and complex schemas :brain:

Fine-Tuner.ai is a seamless plug-and-play AI interface built on LangChain framework! :rocket:

:electric_plug: Link to the plugin

Continuing the discussion from :rocket: Fine-Tuner.ai - Create Custom AI Models:

1 Like

:loudspeaker:"Building Fine-Tuned Chatbots with No-code" tutorial on YouTube :eyes::fire:

:link: https://youtube.com/watch?v=aCW8NgyUfb0

Plus check out our docs for:

:white_check_mark: Slide deck of the tutorial
:white_check_mark: Link to Bubble editor with everything already set up

Link to docs :point_down:
https://finetuner.gitbook.io/fine-tuner.ai/fundamentals/tutorials

All resources :100:% for free!:rocket::gift:

Happy weekend & happy fine-tuning! :raised_hands:

Continuing the discussion from :rocket: Fine-Tuner.ai - Create Custom AI Models:

1 Like

Hi @albert.astabatsyan1 :wave: This is really incredible. I’m very intrigued by the ability to create my own models programmatically via bubble and then supplement them train them on a custom knowledge base. That said, in looking over your documentation, I had a couple (probably ignorant) questions: If using the OpenAI base models, how would one specify a system message or provide context when interacting? By way of context, I’d like to be able to provide an initial system prompt to the chatbot to frame a conversation with an end user, but it’s not clear to me how to do that.