Create README.md

README.md

@@ -0,0 +1,92 @@
+---
+license: apache-2.0
+tags:
+- Flux
+- FluxPipeline
+- text-to-image
+- flux schnell
+- image-generation
+- flux-diffusers
+- diffusers
+- photo
+- realism
+pipeline_tag: text-to-image
+library_name: diffusers
+emoji: 🔜
+language:
+- en
+base_model: black-forest-labs/FLUX.1-schnell
+instance_prompt: HST autochrome photo
+
+widget:
+  - text: hst style photo of a dark-haired woman in a reading room
+    output:
+      url: hstiv9.png
+  - text: hst style autochrome vintage color photo of Rosa Luxemburg
+    output:
+      url: hstiv3.png
+---
+
+# **Historic Color Soon® V.2 (by Soon®)**
+The second Flux-based & open-licensed full-model checkpoint in our **HSToric Color** series.<br>
+
+Trained on HD scans of early color photos (circa *1900s-1910s*) by **Sergey Prokudin-Gorsky**, who traveled and photographed widely in those years whilst perfecting implementations of a pioneering 3-color-composite photography technique.<br>  
+
+**This model is aimed at being useful for**:<br>
+- Quality generation at a low step-count (2 to 8, for most scenarios), with 4-step inference at around 768x768 routinely producing photorealistic outputs at a quality plausibly preferrable to that of Flux v.1 Dev.  <br>
+- Producing realistic images reminiscent of color film analog photography, exhibiting parallels to a broad spectrum of iconic instrumentalities and visual paradigms, from Autochrome-to-Kodachrome-to-Fujifilm-and-beyond. <br>
+- Producing visuals with a vaguely "historical" or "lived-in" aesthetic character, striking chromaticity and luminosity dynamics, as well as textural/anatomical/skin details more reliably lifelike than other models at a comparable step-count/resource expenditure. <br>
+- Extending realism options under an unrevokable commercial license. <br>
+
+<Gallery />
+
+## Testing Space:
+You may try out the **V2** checkpoint at [one of our LoRA gallery spaces](https://huggingface.co/spaces/AlekseyCalvin/soonfactory4), along with this LoRA (and many others)!<br> 
+
+## Bit of Model History + TOOL SHARES:
+[Historic Color Soon® V.1](https://huggingface.co/AlekseyCalvin/HistoricColorSoonr_Schnell) was fine-tuned by us from [HumbleMikey](https://civitai.com/user/humblemikey)'s [Pixelwave Schnell V.1](https://huggingface.co/mikeyandfriends/PixelWave_FLUX.1-schnell_01/) model which, in its turn, is a generalized base checkpoint trained from [FLUX.1-schnell](https://huggingface.co/black-forest-labs/FLUX.1-schnell) by **Black Forest Labs**, consolidating (in comparison w/vanilla-base-**Schnell**) further inference speed improvements (more reliable results at 2-3 steps), whilst raising the overall quality and consistency standards across most aesthetic categories and at every step.<br>
+This version, **Historic Color Soon® V.2** was created through merging into **V.1** a handful of LoRAs trained by us on the (fairly narrow) available range of realistic Flux checkpoint models that are exclusively **Schnell**-derived, so as to stay within the fairly open **Apache 2.0** licensing domain (which was among our reasons to do all this in the first place).<br> 
+
+**Historic Color Soon® V.1** is available [here](https://huggingface.co/AlekseyCalvin/HistoricColorSoonr_Schnell) in both **Safetensors** (fp8) & **Diffusers** formats.<br> 
+
+To fine-tune **Flux**, try the dedicated [Flux Training Notebook by Ostris](https://github.com/ostris/ai-toolkit/blob/main/notebooks/FLUX_1_schnell_LoRA_Training.ipynb).<br>  
+**Ostris**' training adapter for **Schnell** is found here: [ostris/FLUX.1-schnell-training-adapter](https://huggingface.co/ostris/FLUX.1-schnell-training-adapter).<br>  
+
+To merge **Flux*** models and LoRAs, use the *'flux_merge_lora.py'* script from the sd3-branch & /networks (subfolder) of [Kohya-ss's sd-scripts git](https://github.com/kohya-ss/sd-scripts/tree/sd3).<br> 
+
+## Bit of Actual History:
+**Prokudin-Gorsky**'s color photography technique would involve three photo-exposures, either simultaneous or sequential, using specialized color-spectrum filters (basically R.B.G.: red, blue, and green), rendering a subject/shot onto glass plates covered with light-emulsive mixture.<br> 
+The photographer's focus on refining the developer and filter quality, in tandem with his incessant and wide-ranging experimentation, and his artful optimizations of glass plates (generally unwieldly, esp. for color, and by the 1910's already becoming outmoded for B&W on-location shoots, though elsewise extra reliable) ultimately led him to produce a color photography oeuvre of much greater fidelity and vividness than achieved by most of his contemporaries.<br>  
+At the same time, the peculiarities of the photographer's method, coupled with his exceptionally hands-on execution thereof, would manifest in a range of idyosyncratic color, light, and motion artifacts common across the resulting prints.<br>  
+Seldom marring the image as a whole, and less grave than the weaknesses of some cp-emerging autochrome techniques, the warm color hazes & flares framing many of **Prokudin-Gorsky**'s prints constitute a kind of ephemeral signature.<br>
+Alongside some of the more subtle chromatic, textural, and (in some measure) figural characteristics of his work, these auras have reliably imprinted themselves into this and other LoRAs and Models within our gallery of fine-tunes for Flux and StableDiffusion3.5, fine-tuned exclusively on non-synthetic (human-made and pre-curated) open-access data from iconic, influential, and/or otherwise compelling historical sources.<br>
+
+We urge you to explore the works of **Prokudin-Gorsky** for yourself, at the wonderfully organized online [archive at this link](https://prokudin-gorsky.org/), featuring many hundreds of high quality downloadable scans of composite color photo prints from the photographer's original glass plate negatives, available at this site alongside relatively recent restorations of a substantial portion of the images. The original glass-plate negatives are currently held at and administrated by the Library of Congress in Washington, DC, USA. <br> 
+
+## Diffusers:<br>
+To use `Historic Color SOON® V.2` with the 🧨 diffusers python library, first install or upgrade diffusers:<br>
+
+```shell
+pip install -U diffusers
+```
+Then you can use `FluxPipeline` to run the model:
+```python
+import torch
+from diffusers import FluxPipeline
+pipe = FluxPipeline.from_pretrained("AlekseyCalvin/HistoricColorSoonr_v2_FluxSchnell_Diffusers", torch_dtype=torch.bfloat16)
+pipe.to("cuda")
+pipe.enable_model_cpu_offload() #save some VRAM by offloading the model to CPU. Remove this if you have enough GPU power
+prompt = "HST style autochrome film photograph portrait of 1910 woman playing poker against a purple feathered dinosaur, the green-eyed woman has moderately blemished skin with visible lines and pores, she smiles, film grain, Kodachrome"
+image = pipe(
+    prompt,
+    guidance_scale=1.2,
+    num_inference_steps=4,
+    max_sequence_length=256,
+    generator=torch.Generator("cpu").manual_seed(0)
+).images[0]
+image.save("hstcolor1.png")
+```
+To learn more check out the [diffusers](https://huggingface.co/docs/diffusers/main/en/api/pipelines/flux) documentation.
+<br>
+
+Lastly, if you're into literature broadly and old modernist poetry specifically, check out our verse translations at [SILVER AGE POETS](https://www.SilverAgePoets.com/the-poets-and-their-stories)!