AlekseyCalvin
commited on
Commit
•
f7d342c
1
Parent(s):
db37476
Update README.md
Browse files
README.md
CHANGED
@@ -45,9 +45,9 @@ widget:
|
|
45 |
<Gallery />
|
46 |
|
47 |
# Mayakovsky Style Soviet Constructivist Posters & Cartoons Flux LoRA(v.1) by SOON®
|
48 |
-
Trained via Ostris' [ai-toolkit](https://replicate.com/ostris/flux-dev-lora-trainer/train) on 50 high-resolution
|
49 |
For this training experiment, we first spent many days rigorously translating the textual elements (slogans, captions, titles, inset poems, speech fragments, etc), with form/signification/rhymes intact, throughout every image subsequently used for training. <br>
|
50 |
-
These textographic elements were, furthermore, re-placed by us into their original visual contexts, using fonts matched to the sources. <br>
|
51 |
We then manually composed highly detailed paragraph-long captions, wherein we detailed both the graphic and the textual content of each piece, its layout, as well as the most intuitive/intended apprehension of each composition. <br>
|
52 |
This version of the resultent LoRA was trained on our custom Schnell-based checkpoint (Historic Color 2), available here, for 3600 steps at a Transformer Learning Rate of .00002, batch 1, ademamix8bit! No synthetic data, zero auto-generated captions! <br>
|
53 |
|
|
|
45 |
<Gallery />
|
46 |
|
47 |
# Mayakovsky Style Soviet Constructivist Posters & Cartoons Flux LoRA(v.1) by SOON®
|
48 |
+
Trained via Ostris' [ai-toolkit](https://replicate.com/ostris/flux-dev-lora-trainer/train) on 50 high-resolution scans of 1910s/1920s posters & artworks by the great Soviet **poet, artist, & Marxist activist Vladimir Mayakovsky**. <br>
|
49 |
For this training experiment, we first spent many days rigorously translating the textual elements (slogans, captions, titles, inset poems, speech fragments, etc), with form/signification/rhymes intact, throughout every image subsequently used for training. <br>
|
50 |
+
These translated textographic elements were, furthermore, re-placed by us into their original visual contexts, using fonts matched up to the sources. <br>
|
51 |
We then manually composed highly detailed paragraph-long captions, wherein we detailed both the graphic and the textual content of each piece, its layout, as well as the most intuitive/intended apprehension of each composition. <br>
|
52 |
This version of the resultent LoRA was trained on our custom Schnell-based checkpoint (Historic Color 2), available here, for 3600 steps at a Transformer Learning Rate of .00002, batch 1, ademamix8bit! No synthetic data, zero auto-generated captions! <br>
|
53 |
|