English

简体中文
日本語

English

简体中文
日本語

Using Local Models ​

VJSP supports running large language models locally on your computer. This offers several advantages:

  • Privacy: Your code and data never leave your machine.
  • Customization: Experiment with different models and configurations.

However, there are also drawbacks to using local models:

  • Resource Requirements: Local models can be resource-intensive, requiring a powerful computer, preferably with a dedicated GPU.
  • Setup Complexity: Setting up a local model can be more complex than using a cloud API.
  • Model Performance: Performance can vary significantly. While some models perform well, they may not match the largest, most advanced cloud models.
  • Feature Limitations: Local models (and many online models) often lack support for advanced features such as prompt caching, tool use, etc.

ⓘ Info

Having multiple configuration files allows you to quickly switch between different AI providers, models, and settings without having to reconfigure everything every time you want to change providers.

How It Works ​

A configuration file can have its own:

  • API Provider

  • API Key and authentication details

  • Model selection

  • Temperature setting to control response randomness

  • Thinking budget

  • Provider-specific settings

Note that available settings vary by provider and model. Each provider offers different configuration options, and even within the same provider, different models may support different parameter ranges or features.

Setting Up a Local Model ​

  1. Go to the application settings and navigate to the Model Configuration page.
  2. In the configuration file management area, click the + button to create a new configuration and fill in the key parameters as follows:
    • API Provider: Select OpenAI Compatible;
    • OpenAI Base URL: Enter the access address of your local model server;
    • API Key: Enter the authentication key for your local model;
    • Model: Select the name of your deployed local model from the list;
    • Other parameters can be adjusted according to your specific needs.
  3. After verifying all parameters, click the Save button to complete the local model configuration.

Switching Configuration Profiles ​

Switch between configuration profiles in two ways:

  1. From the Settings panel: Select a different profile from the dropdown menu

  2. During a chat: Access the API configuration dropdown in the chat interface

Pinning and Sorting Configurations ​

The API configuration dropdown now supports pinning your favorite configurations for faster access:

  1. Hover over any configuration in the dropdown to reveal the pin icon
  2. Click the pin icon to add the configuration to your pinned list
  3. Pinned configurations appear at the top of the dropdown, sorted alphabetically
  4. Unpinned configurations appear below a separator, also sorted alphabetically
  5. You can click the same icon again to unpin a configuration

This feature makes it easier to navigate between frequently used configurations, especially if you have many.

Editing and Deleting Configurations ​

  • Select a configuration in Settings to modify any of its settings
  • Click the pencil icon to rename a configuration
  • Click the trash icon to delete a configuration (you cannot delete the last remaining configuration)

Security Notes ​

API keys are securely stored in the IDE's Secret Storage and are never exposed in plain text.