📄️ 1. Connect to a Server
The Transformer Lab GUI needs to connect to a Transformer Lab Server in order to get stated. If you are using a Local Connection, you can follow the steps on the Local Connection tab and skip this step.
📄️ 2. Start an Experiment
To start your journey in Transformer Lab, the first step is to create an Experiment.
📄️ 3. Load and Run a Model
Introduction
📄️ 4. Interact with a Model 💬
Once you have loaded a model for inference, and it is shown in the currently running model bar in the header, you can start interacting with it.
📄️ 5. Edit Prompts
Transformer Lab allows you to edit the System prompt both from the Prompts tab and from the Interact->Chat pages.
📄️ 6. Work with Training Data
Download Data
📄️ 7. Evaluate Models
Evaluation in Transformer Lab is implemented very simply in the current version.
📄️ 8. Train and Finetune
<img src={require('./img/training-screen.png').default} alt="Training" width="700
📄️ 9. Publish a Model
Exporting and publishing should come soon. Feel free to help contribute.