Update README.md
Browse files
README.md
CHANGED
|
@@ -24,16 +24,62 @@ base_model: black-forest-labs/FLUX.1-dev
|
|
| 24 |
instance_prompt: dalle-mix
|
| 25 |
license: creativeml-openrail-m
|
| 26 |
---
|
| 27 |
-
# Flux-Dalle-Mix-LoRA
|
| 28 |
|
| 29 |
<Gallery />
|
| 30 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 31 |
|
| 32 |
## Trigger words
|
| 33 |
|
| 34 |
You should use `dalle-mix` to trigger the image generation.
|
| 35 |
|
| 36 |
-
|
| 37 |
## Download model
|
| 38 |
|
| 39 |
Weights for this model are available in Safetensors format.
|
|
|
|
| 24 |
instance_prompt: dalle-mix
|
| 25 |
license: creativeml-openrail-m
|
| 26 |
---
|
| 27 |
+
# Flux-Dalle-Mix-LoRA [ Experimental Version ]
|
| 28 |
|
| 29 |
<Gallery />
|
| 30 |
|
| 31 |
+
- Hosted Here🧨: https://huggingface.co/spaces/prithivMLmods/FLUX-LoRA-DLC
|
| 32 |
+
|
| 33 |
+
**The model is still in the training phase. This is not the final version and may contain artifacts and perform poorly in some cases.**
|
| 34 |
+
|
| 35 |
+
## Model description
|
| 36 |
+
|
| 37 |
+
**prithivMLmods/Flux-Dalle-Mix-LoRA**
|
| 38 |
+
|
| 39 |
+
Image Processing Parameters
|
| 40 |
+
|
| 41 |
+
| Parameter | Value | Parameter | Value |
|
| 42 |
+
|---------------------------|--------|---------------------------|--------|
|
| 43 |
+
| LR Scheduler | constant | Noise Offset | 0.03 |
|
| 44 |
+
| Optimizer | AdamW | Multires Noise Discount | 0.1 |
|
| 45 |
+
| Network Dim | 64 | Multires Noise Iterations | 10 |
|
| 46 |
+
| Network Alpha | 32 | Repeat & Steps | 25 & 3700|
|
| 47 |
+
| Epoch | 15 | Save Every N Epochs | 1 |
|
| 48 |
+
|
| 49 |
+
Labeling: florence2-en(natural language & English)
|
| 50 |
+
|
| 51 |
+
Total Images Used for Training : 44 [ Hi -Res ]
|
| 52 |
+
|
| 53 |
+
## Best Dimensions
|
| 54 |
+
|
| 55 |
+
- 1024 x 1024 (Default)
|
| 56 |
+
|
| 57 |
+
|
| 58 |
+
|
| 59 |
+
| Description |
|
| 60 |
+
|--------------------|
|
| 61 |
+
| dalle-mix, Caricatures pencil drawing, cute woman, mid-20's, color pencils, in the style of Pixar, hilariously funny, exaggerated features, white background --chaos 50 --ar 1:2 --stylize 750 |
|
| 62 |
+
|
| 63 |
+
## Setting Up
|
| 64 |
+
```
|
| 65 |
+
import torch
|
| 66 |
+
from pipelines import DiffusionPipeline
|
| 67 |
+
|
| 68 |
+
base_model = "black-forest-labs/FLUX.1-dev"
|
| 69 |
+
pipe = DiffusionPipeline.from_pretrained(base_model, torch_dtype=torch.bfloat16)
|
| 70 |
+
|
| 71 |
+
lora_repo = "prithivMLmods/Flux-Dalle-Mix-LoRA"
|
| 72 |
+
trigger_word = "dalle-mix"
|
| 73 |
+
pipe.load_lora_weights(lora_repo)
|
| 74 |
+
|
| 75 |
+
device = torch.device("cuda")
|
| 76 |
+
pipe.to(device)
|
| 77 |
+
```
|
| 78 |
|
| 79 |
## Trigger words
|
| 80 |
|
| 81 |
You should use `dalle-mix` to trigger the image generation.
|
| 82 |
|
|
|
|
| 83 |
## Download model
|
| 84 |
|
| 85 |
Weights for this model are available in Safetensors format.
|