Commit
•
2c8823a
1
Parent(s):
309ef78
Update README.md
Browse files
README.md
CHANGED
|
@@ -9,13 +9,23 @@ tags:
|
|
| 9 |
This is the fine-tuned Stable Diffusion model trained on high resolution 3D artworks.
|
| 10 |
Use the tokens **_redshift style_** in your prompts for the effect.
|
| 11 |
|
| 12 |
-
**
|
|
|
|
|
|
|
| 13 |
[](https://patreon.com/user?u=79196446)
|
| 14 |
|
| 15 |
**Characters rendered with the model:**
|
| 16 |
-
![Videogame Samples](https://huggingface.co/nitrosocke/
|
| 17 |
**Cars and Landscapes rendered with the model:**
|
| 18 |
-
![Misc. Samples](https://huggingface.co/nitrosocke/
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 19 |
|
| 20 |
### 🧨 Diffusers
|
| 21 |
|
|
@@ -28,25 +38,17 @@ You can also export the model to [ONNX](https://huggingface.co/docs/diffusers/op
|
|
| 28 |
from diffusers import StableDiffusionPipeline
|
| 29 |
import torch
|
| 30 |
|
| 31 |
-
model_id = "nitrosocke/
|
| 32 |
pipe = StableDiffusionPipeline.from_pretrained(model_id, torch_dtype=torch.float16)
|
| 33 |
pipe = pipe.to("cuda")
|
| 34 |
|
| 35 |
-
prompt = "
|
| 36 |
image = pipe(prompt).images[0]
|
| 37 |
|
| 38 |
image.save("./magical_princess.png")
|
| 39 |
```
|
| 40 |
|
| 41 |
-
|
| 42 |
-
**classic disney style helen mirren as a queen**
|
| 43 |
-
_Steps: 30, Sampler: Euler a, CFG scale: 7, Seed: 3496225274, Size: 512x704_
|
| 44 |
-
|
| 45 |
-
#### Prompt and settings for the Ford Model T:
|
| 46 |
-
**classic disney style Ford Model T - Negative prompt: person**
|
| 47 |
-
_Steps: 20, Sampler: DPM2 Karras, CFG scale: 7, Seed: 4817981, Size: 704x512_
|
| 48 |
-
|
| 49 |
-
This model was trained using the diffusers based dreambooth training by ShivamShrirao using prior-preservation loss and the _train-text-encoder_ flag in 9.000 steps.
|
| 50 |
|
| 51 |
## License
|
| 52 |
|
|
|
|
| 9 |
This is the fine-tuned Stable Diffusion model trained on high resolution 3D artworks.
|
| 10 |
Use the tokens **_redshift style_** in your prompts for the effect.
|
| 11 |
|
| 12 |
+
**The name:** I used Cinema4D for a very long time as my go-to modeling software and always liked the redshift render it came with. That is why I was very sad to see the bad results base SD has connected with its token. This is my attempt at fixing that and showing my passion for this render engine.
|
| 13 |
+
|
| 14 |
+
**If you enjoy my work and want to test new models before release, please consider supporting me**
|
| 15 |
[](https://patreon.com/user?u=79196446)
|
| 16 |
|
| 17 |
**Characters rendered with the model:**
|
| 18 |
+
|
| 19 |
**Cars and Landscapes rendered with the model:**
|
| 20 |
+
|
| 21 |
+
|
| 22 |
+
#### Prompt and settings for Tony Stark:
|
| 23 |
+
**(redshift style) robert downey jr as ironman Negative prompt: glasses helmet**
|
| 24 |
+
_Steps: 40, Sampler: DPM2 Karras, CFG scale: 7, Seed: 908018284, Size: 512x704_
|
| 25 |
+
|
| 26 |
+
#### Prompt and settings for the Ford Mustang:
|
| 27 |
+
**redshift style Ford Mustang**
|
| 28 |
+
_Steps: 20, Sampler: DPM2 Karras, CFG scale: 7, Seed: 579593863, Size: 704x512_
|
| 29 |
|
| 30 |
### 🧨 Diffusers
|
| 31 |
|
|
|
|
| 38 |
from diffusers import StableDiffusionPipeline
|
| 39 |
import torch
|
| 40 |
|
| 41 |
+
model_id = "nitrosocke/redshift-diffusion"
|
| 42 |
pipe = StableDiffusionPipeline.from_pretrained(model_id, torch_dtype=torch.float16)
|
| 43 |
pipe = pipe.to("cuda")
|
| 44 |
|
| 45 |
+
prompt = "redshift style magical princess with golden hair"
|
| 46 |
image = pipe(prompt).images[0]
|
| 47 |
|
| 48 |
image.save("./magical_princess.png")
|
| 49 |
```
|
| 50 |
|
| 51 |
+
This model was trained using the diffusers based dreambooth training by ShivamShrirao using prior-preservation loss and the _train-text-encoder_ flag in 11.000 steps.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 52 |
|
| 53 |
## License
|
| 54 |
|