WEEK 8 · LESSON 10 · TOPIC 4 OF 6 · Practical Application

Practical Application

Lesson progress

0% Complete

SECTION 1 OF 1

In this topic · 1 sections
  1. Overview

Overview

Build a consistent-face generation:

  1. Complete the InstantID installation layers listed above.
  2. Start from a working text-to-image graph.
  3. Add Apply InstantID between the prompt/model stage and the KSampler.
  4. Load a clear reference-face image.
  5. Add ControlNet Loader and select the installed InstantID ControlNet model.
  6. Add InstantID Face Analysis and select CUDA when using the CUDA GPU configuration.
  7. Add InstantID Model Loader and select the installed IP-Adapter file.
  8. Connect the modified model, positive conditioning, and negative conditioning back to the KSampler.
  9. Use a portrait-oriented latent size such as 768 by 1024 for the example.
  10. Enter a scene prompt and negative terms.
  11. Queue the workflow and inspect identity fidelity.
  12. Adjust Weight to change face influence.
  13. Adjust Start and End to change how long identity conditioning operates during denoising.
  14. Change the visual style in the positive prompt while retaining the same reference face to test character consistency.

The same node-based reasoning carries into newer checkpoints. AuraFlow uses the same basic text-to-image workflow with the checkpoint changed to an AuraFlow SafeTensors file. Flux uses a similar pipeline but requires changes to the model, CLIP loader, and KSampler configuration. SDXL is used throughout most of this module because its ecosystem is more mature, with a broad selection of models, LoRAs, plugins, ControlNet tools, and established workflows.