Update README.md

README.md

@@ -1,10 +1,12 @@
 ---
 license: openrail++
-base_model: stabilityai/stable-diffusion-xl-base-1.0
+base_model:
+- stabilityai/stable-diffusion-xl-base-1.0
+- cyberdelia/CyberRealisticXL
 tags:
-  - stable-diffusion-xl
-  - text-to-image
-  - gguf
+- stable-diffusion-xl
+- text-to-image
+- gguf
 ---
 
 # CyberRealistic XL v9.0 — GGUF
@@ -53,7 +55,7 @@ All three images were generated using the same prompt, seed, and sampler setting
 |:---:|:---:|:---:|
 | <small>**Full checkpoint — F16**<br>Standard SDXL checkpoint with the VAE baked in at F16. Reference output.</small> | <small>**GGUF F16 + VAE extracted from checkpoint**<br>UNet in GGUF F16, VAE and CLIP extracted from the same checkpoint in F16. Pixel-perfect identical to the reference.</small> | <small>**GGUF F16 + [madebyollin VAE fp16-fix](https://huggingface.co/madebyollin/sdxl-vae-fp16-fix) — F16**<br>Same as above but with madebyollin's VAE (originally FP32, converted to F16).</small> |
 
-The following table shows the difference between some of the K-quants available with `llama-quantize`.
+The following table shows the difference between some of the K-quants available with `llama-quantize` from [llama.cpp](https://github.com/ggerganov/llama.cpp) using [city96's diffusion model patch](https://github.com/city96/ComfyUI-GGUF/blob/main/tools/README.md).
 
 | ![Q8_0](images/Q8_0.png) | ![Q6_K](images/Q6_K.png) | ![Q4_K_M](images/Q4_K_M.png) |
 |:---:|:---:|:---:|