vllm-omni/tests/diffusion/quantization/artifacts/fp8_z_image_quality/README.md

# FP8 Z-Image Quality Artifact

Generated by:

```bash
CUDA_VISIBLE_DEVICES=0 \
VLLM_OMNI_QUALITY_OUTPUT_DIR=$PWD/tests/diffusion/quantization/artifacts/fp8_z_image_quality \
.venv/bin/python -m pytest \
  tests/diffusion/quantization/test_quantization_quality.py::test_quantization_quality[fp8_z_image] \
  -s -v -m ""
```

Quantized config:

```python
{"method": "fp8", "ignored_layers": ["img_mlp"]}
```

Result: failed with the current `max_lpips=0.15` threshold.

| Metric | Value |
|--------|-------|
| LPIPS | 0.250578 |
| PSNR | 17.042153 dB |
| MAE | 0.076899 |

## Comparison

![BF16 baseline vs FP8 quantized](fp8_z_image_comparison.png)

## Individual Outputs

![BF16 baseline](fp8_z_image_baseline.png)

![FP8 quantized](fp8_z_image_quantized.png)