Gerold Meisinger
commited on
Commit
·
0b3513f
1
Parent(s):
be8d5ad
blub
Browse files
README.md
CHANGED
|
@@ -10,36 +10,33 @@ Based on my GitHub monologs at [Edge Drawing - a Canny alternative](https://gith
|
|
| 10 |
|
| 11 |
Controls image generation by edge maps generated with [EdgeDrawing Parameter-Free](https://github.com/CihanTopal/ED_Lib).
|
| 12 |
|
| 13 |
-
For usage see the model page on [Civitai.com]().
|
| 14 |
-
|
| 15 |
-
# Image dataset
|
| 16 |
-
|
| 17 |
-
* [laion2B-en aesthetics>=6.5 dataset](https://huggingface.co/datasets/ChristophSchuhmann/improved_aesthetics_6.5plus)
|
| 18 |
-
* `--min_image_size 512 --max_aspect_ratio 2 --resize_mode="center_crop" --image_size 512`
|
| 19 |
-
* resulting in 180k images
|
| 20 |
-
|
| 21 |
-
Original
|
| 22 |
|
|
|
|
| 23 |
|
| 24 |
|
| 25 |
|
| 26 |
-
EdgeDrawing Parameter-Free
|
| 27 |
-
|
| 28 |
|
| 29 |
|
| 30 |
|
| 31 |
-
Canndy Edge Detection (default in Automatic1111)
|
| 32 |
-
|
| 33 |
|
| 34 |
|
| 35 |
|
| 36 |
-
Example
|
|
|
|
| 37 |
sampler=UniPC steps=20 cfg=7.5 seed=0 batch=9 model: v1-5-pruned-emaonly.safetensors cherry-picked: 1/9
|
| 38 |
prompt: _a detailed high-quality professional photo of swedish woman standing in front of a mirror, dark brown hair, white hat with purple feather_
|
| 39 |
|
| 40 |
-
|
| 41 |
|
| 42 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 43 |
# Training
|
| 44 |
|
| 45 |
```
|
|
@@ -106,3 +103,8 @@ Update: bug in algorithm produces too sparse images on default, see https://gith
|
|
| 106 |
additional arguments: `--proportion_empty_prompts=0.5`
|
| 107 |
Trained for 40000 steps with default settings
|
| 108 |
=> empty prompts were probably too excessive
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 10 |
|
| 11 |
Controls image generation by edge maps generated with [EdgeDrawing Parameter-Free](https://github.com/CihanTopal/ED_Lib).
|
| 12 |
|
| 13 |
+
For usage see the model page on [Civitai.com](https://civitai.com/models/149740).
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 14 |
|
| 15 |
+
**Original**
|
| 16 |
|
| 17 |
|
| 18 |
|
| 19 |
+
**EdgeDrawing Parameter-Free**
|
|
|
|
| 20 |
|
| 21 |
|
| 22 |
|
| 23 |
+
**Canndy Edge Detection (default in Automatic1111)**
|
|
|
|
| 24 |
|
| 25 |
|
| 26 |
|
| 27 |
+
**Example**
|
| 28 |
+
|
| 29 |
sampler=UniPC steps=20 cfg=7.5 seed=0 batch=9 model: v1-5-pruned-emaonly.safetensors cherry-picked: 1/9
|
| 30 |
prompt: _a detailed high-quality professional photo of swedish woman standing in front of a mirror, dark brown hair, white hat with purple feather_
|
| 31 |
|
|
|
|
| 32 |
|
| 33 |
|
| 34 |
+
# Image dataset
|
| 35 |
+
|
| 36 |
+
* [laion2B-en aesthetics>=6.5 dataset](https://huggingface.co/datasets/ChristophSchuhmann/improved_aesthetics_6.5plus)
|
| 37 |
+
* `--min_image_size 512 --max_aspect_ratio 2 --resize_mode="center_crop" --image_size 512`
|
| 38 |
+
* resulting in 180k images
|
| 39 |
+
|
| 40 |
# Training
|
| 41 |
|
| 42 |
```
|
|
|
|
| 103 |
additional arguments: `--proportion_empty_prompts=0.5`
|
| 104 |
Trained for 40000 steps with default settings
|
| 105 |
=> empty prompts were probably too excessive
|
| 106 |
+
|
| 107 |
+
# Question and answers
|
| 108 |
+
|
| 109 |
+
**Q: What's the point of another edge control net anyway?**
|
| 110 |
+
A: 🤷
|