ArtificialGuyBR

Home / Blog / Training LoRAs

Fish Eye Lens LoRA: Add Fisheye Distortion to Any Image

Apply fisheye camera effect with barrel distortion using this LoRA for Flux, Qwen, and ZImageTurbo. Trigger words, settings, and examples.

1 sources cited Training LoRAs

On this page

  1. What This LoRA Does
  2. Base Models and Compatibility
  3. How to Use
  4. Recommended Settings
  5. License
  6. Common Errors and Fixes
  7. FAQ
  8. Sources

Fish Eye Lens LoRA: Add Fisheye Distortion to Any Image

TL;DR

What This LoRA Does

The Fish Eye Lens LoRA applies a fisheye camera effect to generated images, producing the characteristic barrel distortion where straight lines curve outward from the center and the field of view expands to roughly 180 degrees. This creates the ultra-wide, immersive perspective associated with fisheye lenses in photography.

The LoRA was trained on a dataset of fisheye lens photographs and generalizes the effect across different subjects — architecture, landscapes, interiors, and portraits all gain the distinctive curved perspective.

Fisheye lens example showing barrel distortion on architecture

Base Models and Compatibility

This LoRA targets three base models:

The multi-base training means you can apply the same fisheye style regardless of which backbone you prefer. Safetensors format ensures broad compatibility.

How to Use

Trigger words: Include Fish Eye Lens. Fish eye camera. in your prompt.

Recommended setup (ComfyUI):

  1. Load your chosen base model (Flux.2 Klein 9B, Qwen Image, or ZImageTurbo)
  2. Add the LoRA with strength 0.8–1.0 (start at 1.0)
  3. Use the trigger phrase in your positive prompt
  4. Generate

Example prompt:

Fish Eye Lens. Fish eye camera. modern architecture building, glass facade, sunny day, ultra wide angle

Fisheye lens example on landscape scene

Parameter | Recommendation

|-----------|----------------

LoRA strength | 0.8–1.0 Base model | Flux.2 Klein 9B, Qwen Image, or ZImageTurbo Sampler | Euler a / DPM++ 2M Karras Steps | 20–30 CFG | 3.5–7.0 (Flux), 5–7 (SDXL/Qwen) The author recommends ComfyUI for best results, but the LoRA works in any interface that supports Safetensors LoRAs (Automatic1111, Forge, ComfyUI, diffusers).

Fisheye lens example on interior scene

License

Apache License 2.0 — free for personal and commercial use with attribution.

Common Errors and Fixes

Error | Cause | Fix

|-------|-------|-----

Effect too weak | LoRA strength too low | Increase to 1.0 Effect too strong / artifacts | LoRA strength too high | Reduce to 0.7–0.8 Base model mismatch | Using unsupported base | Use Flux.2 Klein 9B, Qwen Image, or ZImageTurbo Trigger not working | Missing trigger phrase | Add Fish Eye Lens. Fish eye camera. exactly

FAQ

Q: Does this work with SDXL or only Flux/Qwen? A: ZImageTurbo is SDXL-based, so yes — it works on SDXL through that base model.

Q: Can I combine this with other style LoRAs? A: Yes, stack at reduced strengths (0.5–0.7 each) to blend styles.

Q: What resolution works best? A: Native resolution of your base model (1024×1024 for SDXL/ZImageTurbo, 1024×1024 for Flux, varies for Qwen).

Sources