Instructions to use strangerzonehf/Dynamic-Silver-Flux with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Diffusers
How to use strangerzonehf/Dynamic-Silver-Flux with Diffusers:
pip install -U diffusers transformers accelerate
import torch from diffusers import DiffusionPipeline # switch to "mps" for apple devices pipe = DiffusionPipeline.from_pretrained("black-forest-labs/FLUX.1-dev", dtype=torch.bfloat16, device_map="cuda") pipe.load_lora_weights("strangerzonehf/Dynamic-Silver-Flux") prompt = "Dynamic Silver Art, Silver Surfer , Captured at eye-level on a digital art print of a mans torso and upper torso. The mans body is a shiny silver metal, and the mans head is adorned with a white headband. His eyes are a piercing blue, and his mouth is slightly open. His hands are clasped in front of his chest. The backdrop is dark, and there are small white stars in the sky, adding a touch of contrast to the scene. " image = pipe(prompt).images[0] - Inference
- Notebooks
- Google Colab
- Kaggle
- Local Apps Settings
- Draw Things
- DiffusionBee
Update README.md
Browse files
README.md
CHANGED
|
@@ -31,6 +31,48 @@ license: apache-2.0
|
|
| 31 |
|
| 32 |
<Gallery />
|
| 33 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 34 |
|
| 35 |
## Trigger words
|
| 36 |
|
|
|
|
| 31 |
|
| 32 |
<Gallery />
|
| 33 |
|
| 34 |
+
# Model description for Dynamic-Silver-Flux
|
| 35 |
+
|
| 36 |
+
Image Processing Parameters
|
| 37 |
+
|
| 38 |
+
| Parameter | Value | Parameter | Value |
|
| 39 |
+
|---------------------------|--------|---------------------------|--------|
|
| 40 |
+
| LR Scheduler | constant | Noise Offset | 0.03 |
|
| 41 |
+
| Optimizer | AdamW | Multires Noise Discount | 0.1 |
|
| 42 |
+
| Network Dim | 64 | Multires Noise Iterations | 10 |
|
| 43 |
+
| Network Alpha | 32 | Repeat & Steps | 30 & 3850 |
|
| 44 |
+
| Epoch | 22 | Save Every N Epochs | 1 |
|
| 45 |
+
|
| 46 |
+
Labeling: florence2-en(natural language & English)
|
| 47 |
+
|
| 48 |
+
Total Images Used for Training : 25
|
| 49 |
+
|
| 50 |
+
## Best Dimensions & Inference
|
| 51 |
+
|
| 52 |
+
| **Dimensions** | **Aspect Ratio** | **Recommendation** |
|
| 53 |
+
|-----------------|------------------|---------------------------|
|
| 54 |
+
| 1280 x 832 | 3:2 | Best |
|
| 55 |
+
| 1024 x 1024 | 1:1 | Default |
|
| 56 |
+
|
| 57 |
+
### Inference Range
|
| 58 |
+
|
| 59 |
+
- **Recommended Inference Steps:** 30-35
|
| 60 |
+
|
| 61 |
+
## Setting Up
|
| 62 |
+
```python
|
| 63 |
+
import torch
|
| 64 |
+
from pipelines import DiffusionPipeline
|
| 65 |
+
|
| 66 |
+
base_model = "black-forest-labs/FLUX.1-dev"
|
| 67 |
+
pipe = DiffusionPipeline.from_pretrained(base_model, torch_dtype=torch.bfloat16)
|
| 68 |
+
|
| 69 |
+
lora_repo = "strangerzonehf/Dynamic-Silver-Flux"
|
| 70 |
+
trigger_word = "Dynamic Silver Art"
|
| 71 |
+
pipe.load_lora_weights(lora_repo)
|
| 72 |
+
|
| 73 |
+
device = torch.device("cuda")
|
| 74 |
+
pipe.to(device)
|
| 75 |
+
```
|
| 76 |
|
| 77 |
## Trigger words
|
| 78 |
|