ernie-image-gen

ERNIE image generation supports both local deployments and cloud endpoints using template-based configurations. You can switch between openai_compatible and native_generate modes without altering your application code.

Yes, ERNIE image generation supports switching between openai_compatible and native_generate modes simply by updating template-based configuration files, enabling seamless transitions between cloud and local deployments without code changes.

The best way to preserve deployment context like model paths and weights across machines is using portable cloud-config templates. These templates ensure consistent ERNIE image generation configurations during cross-machine portability.

You perform a health check on an ERNIE image generation endpoint by running the health_check script against your configured cloud or local service. This verifies endpoint availability before you attempt to generate images or log artifacts.

Yes, ERNIE image generation automatically logs prompts, images, and metadata as artifacts. It consolidates these outputs into designated artifact directories to preserve the full deployment context and generation history.