Update README.md
Browse files
README.md
CHANGED
@@ -20,4 +20,22 @@ tags:
|
|
20 |
|
21 |
- Text encoder (`text_encoder.`): about 11 GB
|
22 |
- Denoiser (`diffusion_model.`): about 10 GB
|
23 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
20 |
|
21 |
- Text encoder (`text_encoder.`): about 11 GB
|
22 |
- Denoiser (`diffusion_model.`): about 10 GB
|
23 |
+
|
24 |
+
## Samples
|
25 |
+
|
26 |
+
|
27 |
+
|`torch.bfloat16` | Quanto Int8 |
|
28 |
+
| - | - |
|
29 |
+
| <img src="./images/sample_bf16_01.jpg" width="400px" /> | <img src="./images/sample_quanto_01.jpg" width="400px" /> |
|
30 |
+
| VRAM 40GB | VRAM 28GB |
|
31 |
+
|
32 |
+
|
33 |
+
<details><summary>Generation parameters</summary>
|
34 |
+
|
35 |
+
- prompt: `""" A photo of a nendoroid figure of hatsune miku holding a sign that says "CogView4" """"`
|
36 |
+
- negative_prompt: `"blurry, low quality, horror"`
|
37 |
+
- height: `1152`
|
38 |
+
- width: `1152`
|
39 |
+
- cfg_scale: `3.5`
|
40 |
+
- num_inference_steps: `20`
|
41 |
+
</details>
|