README.md

---
license: openrail
---
---
license: openrail
---
The embedding were trained using A1111 TI for the 768px Stable Diffusion v2.0 model. 
The embedding should work on any model that uses SD v2.0 as a base.

**TungstenDispo (v1)**

The TungstenDispo embedding were trained for 1000 epochs with a gradient batch size of 50.
A total of ~100 training images of tungsten photographs taken with CineStill 800T were used. The split was around 50/50 people landscapes.

The effect isn't quite the tungsten photo effect I was going for, but creates very nice, artistic portraits of people. For some of the people, I used SoCalGuitarist's [Negative FaceLift](https://civitai.com/models/2385/socalguitarists-magic-facelift-negative-embedding-for-model-2x-fix-yo-ugly-faces) as a negative embedding.
I used it on 0.3 strength, and it seems like it makes the eyes slightly less wonky. Unclear extent of effect.

Landscapes haven't been experimented with much and are WIP.

<div align="center">
<img src="https://huggingface.co/fzbuzz/TungstenDispo-embedding-sd-v2-1/resolve/main/00000.png">
</div>

<div align="center">
<img src="https://huggingface.co/fzbuzz/TungstenDispo-embedding-sd-v2-1/resolve/main/00001.png">
</div>

<div align="center">
<img src="https://huggingface.co/fzbuzz/TungstenDispo-embedding-sd-v2-1/resolve/main/00002.png">
</div>

<div align="center">
<img src="https://huggingface.co/fzbuzz/TungstenDispo-embedding-sd-v2-1/resolve/main/00003.png">
</div>

---
**Workflow for Above Pictures**

Sampler: Euler-A, 20 Steps, CFG: 7.0. Slightly cherry-picked for best pictures.
900x768 -> 4x LDSR upscaled

Negative Prompt for all images (Not entirely sure if all of them matter, but does help a bit):

> (Neg_Facelift768:0.3), (blur:0.3), (cropped:1.3), (ugly:1.3), (bad anatomy:1.2), (disfigured:1.1), (deformed:1.1), (bad proportions:1.3), (extra limbs:1.2), (missing fingers:1.2), (extra fingers:1.2), (out of frame:1.3), (makeup:1.1)

Positive Prompts:

First Image:

>TungstenDispo, photoshoot of a asian female model with white hair, in a dark room, (closeup:0.2)

Second Image:

>(TungstenDispo:1.3), photoshoot of a (model:0.7), posed, in a dark room, highly detailed, (closeup:0.2), (skin pores:0.5)

Third Image:

>(TungstenDispo:1.2), photoshoot of a model, posed, in a dark room, (closeup:0.2)

Fourth Image:

>TungstenDispo, photoshoot of a male model, posed, in a dark room, (closeup:0.2)

**Usage for A1111 WebUI**

Download the TungstenDispo.pt file and put in embeddings/. Prepend "TungstenDispo" at start of prompt. 


**More People Samples w/out exact workflow**

Pretty much the same, only changed up subject a little + weights.

<div align="center">
<img src="https://huggingface.co/fzbuzz/TungstenDispo-embedding-sd-v2-1/resolve/main/people_sample.png">
</div>


**Landscape Demo:**

Honestly did not turn out as good as anticipated. It really loves neon signs. Stay tuned for more exploration on landscapes :p

Sampler: DDIM, 80 Steps, CFG: 7.0. Slightly cherry-picked for best pictures.
768x1024
Positive Prompt (bolded part altered for image)
>(TungstenStyle:1.5), cinematic shot of a (empty:1.5) **symmetric glass foyer at night with light faintly shining through the windows**, highly detailed

Negative Prompt:
>billboards, signs, words, (blur:0.2)

<div align="center">
<img src="https://huggingface.co/fzbuzz/TungstenDispo-embedding-sd-v2-1/resolve/main/background_sample.png">
</div>