Update README.md
Browse files
README.md
CHANGED
|
@@ -5,10 +5,15 @@ tags:
|
|
| 5 |
- diffusers
|
| 6 |
- unconditional-image-generation
|
| 7 |
- diffusion-models-class
|
|
|
|
|
|
|
| 8 |
---
|
| 9 |
|
| 10 |
# This model is a fine-tuned diffusion model for unconditional image generation of animefaces.
|
| 11 |
|
|
|
|
|
|
|
|
|
|
| 12 |
## Usage
|
| 13 |
|
| 14 |
```python
|
|
@@ -17,4 +22,4 @@ from diffusers import DDPMPipeline
|
|
| 17 |
pipeline = DDPMPipeline.from_pretrained('Pie31415/dm_anime')
|
| 18 |
image = pipeline().images[0]
|
| 19 |
image
|
| 20 |
-
```
|
|
|
|
| 5 |
- diffusers
|
| 6 |
- unconditional-image-generation
|
| 7 |
- diffusion-models-class
|
| 8 |
+
datasets:
|
| 9 |
+
- huggan/selfie2anime
|
| 10 |
---
|
| 11 |
|
| 12 |
# This model is a fine-tuned diffusion model for unconditional image generation of animefaces.
|
| 13 |
|
| 14 |
+
Even after fine-tuning the diffusion model for 10 epochs the generated images are still cursed... 💀. Maybe more epochs would help?
|
| 15 |
+
|
| 16 |
+
|
| 17 |
## Usage
|
| 18 |
|
| 19 |
```python
|
|
|
|
| 22 |
pipeline = DDPMPipeline.from_pretrained('Pie31415/dm_anime')
|
| 23 |
image = pipeline().images[0]
|
| 24 |
image
|
| 25 |
+
```
|