Update README.md

README.md

@@ -16,21 +16,21 @@ instance_prompt: >-
   illustration
 widget:
 - text: >-
-    Somov style art, artwork by Konstantin SomovSomov style art featuring
+    (step 3600) Somov style art, artwork by Konstantin SomovSomov style art featuring
     artwork by Konstantin Somov with a carnival scene taking place in the 18th
     century, a man and a woman in Italian masks, in a garden, detailed
     illustration!
   output:
     url: images/example_tnkmv0i1l.png
 - text: >-
-    Somov style art, artwork by Konstantin Somov, detailed carnivalesque
+    (step 3600) Somov style art, artwork by Konstantin Somov, detailed carnivalesque
     illustration of a tiny drugged God inside of a wooden pillbox getting
     impaled by a giant red-robed woman wielding a huge sharp needle pin,
     indoors, in a hotel room in Vienna
   output:
     url: images/example_by5i130ad.png
 - text: >-
-    Somov style art, artwork by Konstantin Somov, detailed carnivalesque
+    (step 3600) Somov style art, artwork by Konstantin Somov, detailed carnivalesque
     illustration of of Arthur Rimbaud, the revolutionary blonde messy haired
     teenage trans man poet delivering a lecture to parlor book club group of
     revolutionary poet cat-person hermaphrodite angels circa the 1870s
@@ -42,9 +42,12 @@ widget:
 ## HISTORIC 𝒜𝓡𝒯𝚒𝕤𝕥 STYLE ℒ𝓸ℜ𝐴𝘴
 ## ||| By SilverAgePoets.com |||
 ![Theatre](https://huggingface.co/AlekseyCalvin/Somov_GreatArtistStyle_FluxLoRA_bySilverAgePoets/resolve/main/Theatre.png)
+*Note: **Variant II** checkpoints of this model were fine-tuned at 64dim64alpha over [AfroMan4Peace](https://civitai.com/user/Afroman4peace)'s [Colossus2.1_Dedistilled](https://civitai.com/models/833086?modelVersionId=996001) + [LongCLIP-GmP-ViT-L-14 by zer0int](https://huggingface.co/zer0int/LongCLIP-GmP-ViT-L-14).* <br> 
+*Key hyperparameters listed below.* <br>
+***Variant I** checkpoints are found [HERE](https://huggingface.co/AlekseyCalvin/Somov_Carnivalesques_V1_FluxLoRA_BySilverAgePoets): fine-tuned at 64dim128alpha and over [DNA_1_618](https://civitai.com/user/DNA_1_618)'s [Genova Apex De-distilled](https://civitai.com/models/954608?modelVersionId=1068773) + [LongCLIP-GmP-ViT-L-14 by zer0int](https://huggingface.co/zer0int/LongCLIP-GmP-ViT-L-14).* <br> 
 
 A Low-Rank Adapter (LoRA) for FLUX-family models... <br>
-Fine-tuned on around 150 artworks inspired by Pierrot's lace tears, Columbina's satin gasps, Harlequin's velvet prances, by the toy moonlight of dreamed-up 18th centuries, and every achingly real rendezvous with fleeting masks of closeness...   <br>
+Fine-tuned on around 150 artworks inspired by Pierrot's lace tears, Columbina's satin gasps, Harlequin's velvet prances, by the toy moonlight of dreamed-up 18th centuries, and every achingly real rendezvous with fleeting masks of closeness...<br>
 And these artworks, of course, were crafted once upon a world by: <br>
 **KONSTANTIN SOMOV** *(1869-1939)*, though representing but one facet, a brightest of dreams and among the deepest, of his manifold eclectic masteries. <br>
 
@@ -55,6 +58,12 @@ Scribble `Somov style art, artwork by Konstantin Somov, detailed carnivalesque i
 
 ![Reverie0](https://huggingface.co/AlekseyCalvin/Somov_GreatArtistStyle_FluxLoRA_bySilverAgePoets/resolve/main/0123.jpeg)
 
+## Hyperparameters:
+Fine-tuned on around 150 images, picked from a broader collection of 300+ Somov art files consolidated by us, and manually pre-processed/restored by us.  <br> 
+Upscaling was used sparingly, applied to only a small portion of the pieces; namely, only to those images which could not be found anywhere online in high quality or/and resolution.<br> 
+Of the upscaled images, we added to the set only those that, after the upscale, did not appear to suffer extreme degradation texture/detail-wise, but seem unambiguously vivified in their roles as conflatable artistic translations apropos respective sources. <br> 
+**Somov the key Hyperparameters:** *64dim/64alpha, Adamw8bit, lr.0003, ema:.99decay, linear timesteps, FluxCFG, resol.768, 1xbatch, on a single L4 via Google Colab Pro. Alas, due to the size of the dataset, we had to limit the resolution to 768. Even 1024 would cause out-of-memory errors, sometimes well into the training, at the tail end of the first epoch.*  <br>
+
 
 ![Reverie1](https://huggingface.co/AlekseyCalvin/Somov_GreatArtistStyle_FluxLoRA_bySilverAgePoets/resolve/main/0223.png)