Update README.md

README.md

@@ -51,6 +51,28 @@ Using [llama.cpp quantize cae9fb4](https://github.com/ggerganov/llama.cpp/commit
 
 - Bravo IQ1_S worse than Alpha?
 - [Latent loss](https://huggingface.co/Eviation/flux-imatrix/blob/main/images/latent_loss.png)
+- [Per layer quantization cost](https://huggingface.co/Eviation/flux-imatrix/blob/main/images/latent_loss.png) from [chrisgoringe/casting_cost](https://github.com/chrisgoringe/mixed-gguf-converter/blob/main/costs/casting_cost.yaml)
+
+
+| Weight name |
+| double_blocks.%d.txt_attn.proj.weight |
+| double_blocks.%d.txt_mod.lin.weight |
+| time_in.in_layer.weight && time_in.out_layer.weight |
+| vector_in.in_layer.weight && vector_in.out_layer.weight |
+| double_blocks.%d.img_mod.lin.weight |
+| double_blocks.%d.txt_attn.qkv.weight |
+| single_blocks.%d.modulation.lin.weight |
+| txt_in.weight |
+| --- |
+| img_in.weight |
+| double_blocks.%d.img_mlp.0.weight && double_blocks.%d.img_mlp.2.weight |
+| double_blocks.%d.img_attn.proj.weight |
+| single_blocks.%d.linear2.weight |
+| final_layer.adaLN_modulation.1.weight && final_layer.linear.weight |
+| double_blocks.%d.txt_mlp.0.weight && double_blocks.%d.txt_mlp.2.weight |
+| double_blocks.%d.img_attn.qkv.weight |
+| single_blocks.%d.linear1.weight |
+| guidance_in.in_layer.weight && guidance_in.out_layer.weight |
 
 # Alpha
 Simple imatrix: 512x512 single image 8/20 steps [city96/flux1-dev-Q3_K_S](https://huggingface.co/city96/FLUX.1-dev-gguf/blob/main/flux1-dev-Q3_K_S.gguf) euler