|
<!--Copyright 2023 The HuggingFace Team. All rights reserved. |
|
|
|
Licensed under the Apache License, Version 2.0 (the "License"); you may not use this file except in compliance with |
|
the License. You may obtain a copy of the License at |
|
|
|
http://www.apache.org/licenses/LICENSE-2.0 |
|
|
|
Unless required by applicable law or agreed to in writing, software distributed under the License is distributed on |
|
an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. See the License for the |
|
specific language governing permissions and limitations under the License. |
|
--> |
|
|
|
# Shap-E |
|
|
|
[[open-in-colab]] |
|
|
|
Shap-Eλ λΉλμ€ κ²μ κ°λ°, μΈν
λ¦¬μ΄ λμμΈ, 건μΆμ μ¬μ©ν μ μλ 3D μμ
μ μμ±νκΈ° μν conditional λͺ¨λΈμ
λλ€. λκ·λͺ¨ 3D μμ
λ°μ΄ν°μ
μ νμ΅λμκ³ , κ° μ€λΈμ νΈμ λ λ§μ λ·°λ₯Ό λ λλ§νκ³ 4K point cloud λμ 16Kλ₯Ό μμ±νλλ‘ νμ²λ¦¬ν©λλ€. Shap-E λͺ¨λΈμ λ λ¨κ³λ‘ νμ΅λ©λλ€: |
|
|
|
1. μΈμ½λκ° 3D μμ
μ ν¬μΈνΈ ν΄λΌμ°λμ λ λλ§λ λ·°λ₯Ό λ°μλ€μ΄κ³ μμ
μ λνλ΄λ implicit functionsμ νλΌλ―Έν°λ₯Ό μΆλ ₯ν©λλ€. |
|
2. μΈμ½λκ° μμ±ν latentsλ₯Ό λ°νμΌλ‘ diffusion λͺ¨λΈμ νλ ¨νμ¬ neural radiance fields(NeRF) λλ textured 3D λ©μλ₯Ό μμ±νμ¬ λ€μ΄μ€νΈλ¦Ό μ ν리μΌμ΄μ
μμ 3D μμ
μ λ μ½κ² λ λλ§νκ³ μ¬μ©ν μ μλλ‘ ν©λλ€. |
|
|
|
μ΄ κ°μ΄λμμλ Shap-Eλ₯Ό μ¬μ©νμ¬ λλ§μ 3D μμ
μ μμ±νλ λ°©λ²μ 보μ
λλ€! |
|
|
|
μμνκΈ° μ μ λ€μ λΌμ΄λΈλ¬λ¦¬κ° μ€μΉλμ΄ μλμ§ νμΈνμΈμ: |
|
|
|
```py |
|
# Colabμμ νμν λΌμ΄λΈλ¬λ¦¬λ₯Ό μ€μΉνκΈ° μν΄ μ£Όμμ μ μΈνμΈμ |
|
#!pip install -q diffusers transformers accelerate trimesh |
|
``` |
|
|
|
## Text-to-3D |
|
|
|
3D κ°μ²΄μ gifλ₯Ό μμ±νλ €λ©΄ ν
μ€νΈ ν둬ννΈλ₯Ό [`ShapEPipeline`]μ μ λ¬ν©λλ€. νμ΄νλΌμΈμ 3D κ°μ²΄λ₯Ό μμ±νλ λ° μ¬μ©λλ μ΄λ―Έμ§ νλ μ 리μ€νΈλ₯Ό μμ±ν©λλ€. |
|
|
|
```py |
|
import torch |
|
from diffusers import ShapEPipeline |
|
|
|
device = torch.device("cuda" if torch.cuda.is_available() else "cpu") |
|
|
|
pipe = ShapEPipeline.from_pretrained("openai/shap-e", torch_dtype=torch.float16, variant="fp16") |
|
pipe = pipe.to(device) |
|
|
|
guidance_scale = 15.0 |
|
prompt = ["A firecracker", "A birthday cupcake"] |
|
|
|
images = pipe( |
|
prompt, |
|
guidance_scale=guidance_scale, |
|
num_inference_steps=64, |
|
frame_size=256, |
|
).images |
|
``` |
|
|
|
μ΄μ [`~utils.export_to_gif`] ν¨μλ₯Ό μ¬μ©νμ¬ μ΄λ―Έμ§ νλ μ 리μ€νΈλ₯Ό 3D κ°μ²΄μ gifλ‘ λ³νν©λλ€. |
|
|
|
```py |
|
from diffusers.utils import export_to_gif |
|
|
|
export_to_gif(images[0], "firecracker_3d.gif") |
|
export_to_gif(images[1], "cake_3d.gif") |
|
``` |
|
|
|
<div class="flex gap-4"> |
|
<div> |
|
<img class="rounded-xl" src="https://huggingface.co/datasets/hf-internal-testing/diffusers-images/resolve/main/shap_e/firecracker_out.gif"/> |
|
<figcaption class="mt-2 text-center text-sm text-gray-500">prompt = "A firecracker"</figcaption> |
|
</div> |
|
<div> |
|
<img class="rounded-xl" src="https://huggingface.co/datasets/hf-internal-testing/diffusers-images/resolve/main/shap_e/cake_out.gif"/> |
|
<figcaption class="mt-2 text-center text-sm text-gray-500">prompt = "A birthday cupcake"</figcaption> |
|
</div> |
|
</div> |
|
|
|
## Image-to-3D |
|
|
|
λ€λ₯Έ μ΄λ―Έμ§λ‘λΆν° 3D κ°μ²΄λ₯Ό μμ±νλ €λ©΄ [`ShapEImg2ImgPipeline`]μ μ¬μ©ν©λλ€. κΈ°μ‘΄ μ΄λ―Έμ§λ₯Ό μ¬μ©νκ±°λ μμ ν μλ‘μ΄ μ΄λ―Έμ§λ₯Ό μμ±ν μ μμ΅λλ€. [Kandinsky 2.1](../api/pipelines/kandinsky) λͺ¨λΈμ μ¬μ©νμ¬ μ μ΄λ―Έμ§λ₯Ό μμ±ν΄ λ³΄κ² μ΅λλ€. |
|
|
|
```py |
|
from diffusers import DiffusionPipeline |
|
import torch |
|
|
|
prior_pipeline = DiffusionPipeline.from_pretrained("kandinsky-community/kandinsky-2-1-prior", torch_dtype=torch.float16, use_safetensors=True).to("cuda") |
|
pipeline = DiffusionPipeline.from_pretrained("kandinsky-community/kandinsky-2-1", torch_dtype=torch.float16, use_safetensors=True).to("cuda") |
|
|
|
prompt = "A cheeseburger, white background" |
|
|
|
image_embeds, negative_image_embeds = prior_pipeline(prompt, guidance_scale=1.0).to_tuple() |
|
image = pipeline( |
|
prompt, |
|
image_embeds=image_embeds, |
|
negative_image_embeds=negative_image_embeds, |
|
).images[0] |
|
|
|
image.save("burger.png") |
|
``` |
|
|
|
μΉμ¦λ²κ±°λ₯Ό [`ShapEImg2ImgPipeline`]μ μ λ¬νμ¬ 3D representationμ μμ±ν©λλ€. |
|
|
|
```py |
|
from PIL import Image |
|
from diffusers import ShapEImg2ImgPipeline |
|
from diffusers.utils import export_to_gif |
|
|
|
pipe = ShapEImg2ImgPipeline.from_pretrained("openai/shap-e-img2img", torch_dtype=torch.float16, variant="fp16").to("cuda") |
|
|
|
guidance_scale = 3.0 |
|
image = Image.open("burger.png").resize((256, 256)) |
|
|
|
images = pipe( |
|
image, |
|
guidance_scale=guidance_scale, |
|
num_inference_steps=64, |
|
frame_size=256, |
|
).images |
|
|
|
gif_path = export_to_gif(images[0], "burger_3d.gif") |
|
``` |
|
|
|
<div class="flex gap-4"> |
|
<div> |
|
<img class="rounded-xl" src="https://huggingface.co/datasets/hf-internal-testing/diffusers-images/resolve/main/shap_e/burger_in.png"/> |
|
<figcaption class="mt-2 text-center text-sm text-gray-500">cheeseburger</figcaption> |
|
</div> |
|
<div> |
|
<img class="rounded-xl" src="https://huggingface.co/datasets/hf-internal-testing/diffusers-images/resolve/main/shap_e/burger_out.gif"/> |
|
<figcaption class="mt-2 text-center text-sm text-gray-500">3D cheeseburger</figcaption> |
|
</div> |
|
</div> |
|
|
|
## λ©μ μμ±νκΈ° |
|
|
|
Shap-Eλ λ€μ΄μ€νΈλ¦Ό μ ν리μΌμ΄μ
μ λ λλ§ν textured λ©μ μΆλ ₯μ μμ±ν μλ μλ μ μ°ν λͺ¨λΈμ
λλ€. μ΄ μμ μμλ π€ Datasets λΌμ΄λΈλ¬λ¦¬μμ [Dataset viewer](https://huggingface.co/docs/hub/datasets-viewer#dataset-preview)λ₯Ό μ¬μ©ν΄ λ©μ μκ°νλ₯Ό μ§μνλ `glb` νμΌλ‘ λ³νν©λλ€. |
|
|
|
`output_type` 맀κ°λ³μλ₯Ό `"mesh"`λ‘ μ§μ ν¨μΌλ‘μ¨ [`ShapEPipeline`]κ³Ό [`ShapEImg2ImgPipeline`] λͺ¨λμ λν λ©μ μΆλ ₯μ μμ±ν μ μμ΅λλ€: |
|
|
|
```py |
|
import torch |
|
from diffusers import ShapEPipeline |
|
|
|
device = torch.device("cuda" if torch.cuda.is_available() else "cpu") |
|
|
|
pipe = ShapEPipeline.from_pretrained("openai/shap-e", torch_dtype=torch.float16, variant="fp16") |
|
pipe = pipe.to(device) |
|
|
|
guidance_scale = 15.0 |
|
prompt = "A birthday cupcake" |
|
|
|
images = pipe(prompt, guidance_scale=guidance_scale, num_inference_steps=64, frame_size=256, output_type="mesh").images |
|
``` |
|
|
|
λ©μ μΆλ ₯μ `ply` νμΌλ‘ μ μ₯νλ €λ©΄ [`~utils.export_to_ply`] ν¨μλ₯Ό μ¬μ©ν©λλ€: |
|
|
|
<Tip> |
|
|
|
μ νμ μΌλ‘ [`~utils.export_to_obj`] ν¨μλ₯Ό μ¬μ©νμ¬ λ©μ μΆλ ₯μ `obj` νμΌλ‘ μ μ₯ν μ μμ΅λλ€. λ€μν νμμΌλ‘ λ©μ μΆλ ₯μ μ μ₯ν μ μμ΄ λ€μ΄μ€νΈλ¦Όμμ λμ± μ μ°νκ² μ¬μ©ν μ μμ΅λλ€! |
|
|
|
</Tip> |
|
|
|
```py |
|
from diffusers.utils import export_to_ply |
|
|
|
ply_path = export_to_ply(images[0], "3d_cake.ply") |
|
print(f"Saved to folder: {ply_path}") |
|
``` |
|
|
|
κ·Έ λ€μ trimesh λΌμ΄λΈλ¬λ¦¬λ₯Ό μ¬μ©νμ¬ `ply` νμΌμ `glb` νμΌλ‘ λ³νν μ μμ΅λλ€: |
|
|
|
```py |
|
import trimesh |
|
|
|
mesh = trimesh.load("3d_cake.ply") |
|
mesh_export = mesh.export("3d_cake.glb", file_type="glb") |
|
``` |
|
|
|
κΈ°λ³Έμ μΌλ‘ λ©μ μΆλ ₯μ μλμͺ½ μμ μ μ΄μ μ΄ λ§μΆ°μ Έ μμ§λ§ νμ λ³νμ μ μ©νμ¬ κΈ°λ³Έ μμ μ λ³κ²½ν μ μμ΅λλ€: |
|
|
|
```py |
|
import trimesh |
|
import numpy as np |
|
|
|
mesh = trimesh.load("3d_cake.ply") |
|
rot = trimesh.transformations.rotation_matrix(-np.pi / 2, [1, 0, 0]) |
|
mesh = mesh.apply_transform(rot) |
|
mesh_export = mesh.export("3d_cake.glb", file_type="glb") |
|
``` |
|
|
|
λ©μ νμΌμ λ°μ΄ν°μ
λ ν¬μ§ν 리μ μ
λ‘λν΄ Dataset viewerλ‘ μκ°ννμΈμ! |
|
|
|
<div class="flex justify-center"> |
|
<img class="rounded-xl" src="https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/diffusers/3D-cake.gif"/> |
|
</div> |
|
|