Quickstart
Getting started with setting up and data tracking.
Introduction to Tracking with FinetuneDB
FinetuneDB enables efficient tracking and analysis of LLM interactions, essential for fine-tuning model performance through data-driven insights.
Step 1: Sign Up and Create an Account
Register on the FinetuneDB website via Google login or email (verify your email to activate your account) and complete the sign-up process.
![Hero Light](https://mintlify.s3-us-west-1.amazonaws.com/finetunedb/images/signup.webp)
![Hero Dark](https://mintlify.s3-us-west-1.amazonaws.com/finetunedb/images/signup.webp)
Step 2: Create a Workspace and Generate an API Key
After logging in, navigate to the dashboard to create a new workspace, give it a custom name such as your project or company. Create a first project, and then generate an API key in the workspace settings for secure connectivity.
![Hero Light](https://mintlify.s3-us-west-1.amazonaws.com/finetunedb/images/apikey.webp)
![Hero Dark](https://mintlify.s3-us-west-1.amazonaws.com/finetunedb/images/apikey.webp)
Step 3: Start Tracking Data
Download and integrate the FinetuneDB SDK into your application, following the setup instructions for your specific environment, and configure it using your new API key to start sending data to your workspace.
With the SDK configured, your application will automatically begin sending data to FinetuneDB.
Step 4: Create Datasets
In the project, use the Log Viewer to filter and select relevant data from production and import the logs to a new or existing dataset. Alternatively, import an already existing dataset into FintuneDB via the upload function or create one manually from scratch.
![Create Dataset](https://mintlify.s3-us-west-1.amazonaws.com/finetunedb/images/add-logs-to-dataset.webp)
![Create Dataset](https://mintlify.s3-us-west-1.amazonaws.com/finetunedb/images/add-logs-to-dataset.webp)
Step 5: Finetune and Compare
Initiate the fine-tuning process by selecting the dataset and base model. Configure training parameters and start the training job. Once the model is trained, compare and evaluate its performance with other models in the Studio.
![Fine-Tune and Compare](https://mintlify.s3-us-west-1.amazonaws.com/finetunedb/images/finetune-compact.webp)
![Fine-Tune and Compare](https://mintlify.s3-us-west-1.amazonaws.com/finetunedb/images/finetune-compact.webp)