Update README.md
Browse files
README.md
CHANGED
@@ -6,18 +6,29 @@ Diffusers formation for mochi-1-preview model.
|
|
6 |
It was create by scripts: https://github.com/huggingface/diffusers/blob/mochi/scripts/convert_mochi_to_diffusers.py
|
7 |
The model can be directly load from pretrained with mochi branch: https://github.com/huggingface/diffusers/tree/mochi
|
8 |
```bash
|
|
|
9 |
from diffusers import MochiPipeline
|
10 |
from diffusers.utils import export_to_video
|
11 |
-
|
|
|
12 |
pipe.to("cuda")
|
13 |
-
prompt = "
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
14 |
frames = pipe(prompt,
|
15 |
-
num_inference_steps=
|
16 |
guidance_scale=4.5,
|
17 |
num_frames=61,
|
18 |
generator=torch.Generator(device="cuda").manual_seed(42),
|
19 |
).frames[0]
|
20 |
-
|
|
|
21 |
```
|
22 |
Some generated results:
|
23 |
<video controls autoplay src="https://cdn-uploads.huggingface.co/production/uploads/62e34a12c9bece303d146af8/Cm3I6kidy2YP5nu3un7XP.mp4"></video>
|
|
|
6 |
It was create by scripts: https://github.com/huggingface/diffusers/blob/mochi/scripts/convert_mochi_to_diffusers.py
|
7 |
The model can be directly load from pretrained with mochi branch: https://github.com/huggingface/diffusers/tree/mochi
|
8 |
```bash
|
9 |
+
import torch
|
10 |
from diffusers import MochiPipeline
|
11 |
from diffusers.utils import export_to_video
|
12 |
+
|
13 |
+
pipe = MochiPipeline.from_pretrained('HanLiii/mochi-1-preview-diffusers', torch_dtype=torch.float16)
|
14 |
pipe.to("cuda")
|
15 |
+
prompt = """
|
16 |
+
A hand with delicate fingers picks up a bright yellow lemon from a wooden bowl
|
17 |
+
filled with lemons and sprigs of mint against a peach-colored background.
|
18 |
+
The hand gently tosses the lemon up and catches it, showcasing its smooth texture.
|
19 |
+
A beige string bag sits beside the bowl, adding a rustic touch to the scene.
|
20 |
+
Additional lemons, one halved, are scattered around the base of the bowl.
|
21 |
+
The even lighting enhances the vibrant colors and creates a fresh,
|
22 |
+
inviting atmosphere.
|
23 |
+
"""
|
24 |
frames = pipe(prompt,
|
25 |
+
num_inference_steps=64,
|
26 |
guidance_scale=4.5,
|
27 |
num_frames=61,
|
28 |
generator=torch.Generator(device="cuda").manual_seed(42),
|
29 |
).frames[0]
|
30 |
+
|
31 |
+
export_to_video(frames, "mochi_test.mp4")
|
32 |
```
|
33 |
Some generated results:
|
34 |
<video controls autoplay src="https://cdn-uploads.huggingface.co/production/uploads/62e34a12c9bece303d146af8/Cm3I6kidy2YP5nu3un7XP.mp4"></video>
|