Shap-E のインストールと動作確認（テキストや画像からの立体生成）（Python，PyTorch を使用）（Windows 上）

python -m pip install -U --ignore-installed pip
python -m pip install -U torch torchvision torchaudio numpy --index-url https://download.pytorch.org/whl/cu118
python -c "import torch; print(torch.__version__, torch.cuda.is_available())"

conda install -y pytorch torchvision torchaudio pytorch-cuda=11.8 cudnn -c pytorch -c nvidia
py -c "import torch; print(torch.__version__, torch.cuda.is_available())"

cd %HOMEPATH%
rmdir /s /q shap-e
git clone --recursive https://github.com/openai/shap-e
cd shap-e
python -m pip install -e .

cd %HOMEPATH%\shap-e
notepad a.py

import torch
import matplotlib.pyplot as plt
from PIL import Image

from shap_e.diffusion.sample import sample_latents
from shap_e.diffusion.gaussian_diffusion import diffusion_from_config
from shap_e.models.download import load_model, load_config
from shap_e.util.notebooks import create_pan_cameras, decode_latent_images, gif_widget

device = torch.device('cuda' if torch.cuda.is_available() else 'cpu')

xm = load_model('transmitter', device=device)
model = load_model('text300M', device=device)
diffusion = diffusion_from_config(load_config('diffusion'))

batch_size = 4
guidance_scale = 15.0
prompt = "a shark"

latents = sample_latents(
    batch_size=batch_size,
    model=model,
    diffusion=diffusion,
    guidance_scale=guidance_scale,
    model_kwargs=dict(texts=[prompt] * batch_size),
    progress=True,
    clip_denoised=True,
    use_fp16=True,
    use_karras=True,
    karras_steps=64,
    sigma_min=1e-3,
    sigma_max=160,
    s_churn=0,
)

render_mode = 'nerf' # you can change this to 'stf'
size = 64 # this is the size of the renders; higher values take longer to render.

cameras = create_pan_cameras(size, device)
for i, latent in enumerate(latents):
    images = decode_latent_images(xm, latent, cameras, rendering_mode=render_mode)
    fig, axes = plt.subplots(nrows=len(images), figsize=(8, 8))
    for ax, image in zip(axes, images):
        ax.imshow(image)
        ax.axis('off')
    plt.tight_layout()
    plt.show()

# Example of saving the latents as meshes.
from shap_e.util.notebooks import decode_latent_mesh

for i, latent in enumerate(latents):
    t = decode_latent_mesh(xm, latent).tri_mesh()
    with open(f'example_mesh_{i}.ply', 'wb') as f:
        t.write_ply(f)
    with open(f'example_mesh_{i}.obj', 'w') as f:
        t.write_obj(f)

python -m pip install PyOpenGL PyOpenGL-accelerate pygame pywavefront

cd %HOMEPATH%\shap-e
notepad d.py

import torch
import matplotlib.pyplot as plt
from PIL import Image

from shap_e.diffusion.sample import sample_latents
from shap_e.diffusion.gaussian_diffusion import diffusion_from_config
from shap_e.models.download import load_model, load_config
from shap_e.util.notebooks import create_pan_cameras, decode_latent_images, gif_widget

print("prompt=")
prompt = input()

device = torch.device('cuda' if torch.cuda.is_available() else 'cpu')

xm = load_model('transmitter', device=device)
model = load_model('text300M', device=device)
diffusion = diffusion_from_config(load_config('diffusion'))

batch_size = 4
guidance_scale = 15.0

latents = sample_latents(
    batch_size=batch_size,
    model=model,
    diffusion=diffusion,
    guidance_scale=guidance_scale,
    model_kwargs=dict(texts=[prompt] * batch_size),
    progress=True,
    clip_denoised=True,
    use_fp16=True,
    use_karras=True,
    karras_steps=64,
    sigma_min=1e-3,
    sigma_max=160,
    s_churn=0,
)

render_mode = 'nerf' # you can change this to 'stf'
size = 64 # this is the size of the renders; higher values take longer to render.

cameras = create_pan_cameras(size, device)
for i, latent in enumerate(latents):
    images = decode_latent_images(xm, latent, cameras, rendering_mode=render_mode)

# Example of saving the latents as meshes.
from shap_e.util.notebooks import decode_latent_mesh

for i, latent in enumerate(latents):
    t = decode_latent_mesh(xm, latent).tri_mesh()
    with open(f'example_mesh_{i}.ply', 'wb') as f:
        t.write_ply(f)
    with open(f'example_mesh_{i}.obj', 'w') as f:
        t.write_obj(f)

import pygame
from pygame.locals import *
from OpenGL.GL import *
from OpenGL.GLUT import *
from OpenGL.GLU import *
from plyfile import PlyData, PlyElement

# PLYファイルを読み込む
import tkinter as tk
from tkinter import filedialog

root = tk.Tk()
root.withdraw()
fpath = filedialog.askopenfilename()

plydata = PlyData.read(fpath)
vertex_data = plydata['vertex'].data
face_data = plydata['face'].data

# 頂点と面をリストとして取得
vertices = [list(elem) for elem in vertex_data]
faces = [face[0] for face in face_data]

# PygameとOpenGLを初期化
pygame.init()
display = (800, 600)
pygame.display.set_mode(display, DOUBLEBUF | OPENGL)
gluPerspective(45, (display[0] / display[1]), 0.1, 50.0)
glTranslatef(0.0, 0.0, -5)

# マウス操作のための変数
rotation_enabled = False
rotation_start = (0, 0)
zoom = 0.0

# メインループ
while True:
    for event in pygame.event.get():
        if event.type == pygame.QUIT:
            pygame.quit()
            quit()
        elif event.type == pygame.MOUSEBUTTONDOWN:
            if event.button == 1:  # 左クリック
                rotation_enabled = True
                rotation_start = pygame.mouse.get_pos()
            elif event.button == 4:  # ホイール回転（上方向）
                zoom = 1
            elif event.button == 5:  # ホイール回転（下方向）
                zoom = -1
        elif event.type == pygame.MOUSEBUTTONUP:
            if event.button == 1:  # 左クリック解除
                rotation_enabled = False
        elif event.type == pygame.MOUSEMOTION:
            if rotation_enabled:
                dx = event.pos[0] - rotation_start[0]
                dy = event.pos[1] - rotation_start[1]
                glRotatef(dx, 0, 1, 0)  # y軸周りに回転
                glRotatef(dy, 1, 0, 0)  # x軸周りに回転
                rotation_start = event.pos

    glClear(GL_COLOR_BUFFER_BIT | GL_DEPTH_BUFFER_BIT)

    # 視点の移動範囲を制限
    max_zoom = 5.0
    min_zoom = -5.0
    zoom = max(min_zoom, min(max_zoom, zoom))
    glTranslatef(0.0, 0.0, -zoom)
    zoom = 0

    glBegin(GL_TRIANGLES)
    for face in faces:
        for vertex_i in face:
            glVertex3fv(vertices[vertex_i][:3])
    glEnd()

    pygame.display.flip()
    pygame.time.wait(10)

Shap-E のインストールと動作確認（テキストや画像からの立体生成）（Python，PyTorch を使用）（Windows 上）

Shap-E

前準備

Git のインストール（Windows 上）

Python のインストール（Windows 上）

Build Tools for Visual Studio 2022，NVIDIA ドライバ，NVIDIA CUDA ツールキット 11.8，NVIDIA cuDNN 8.6 のインストール（Windows 上）

PyTorch のインストール（Windows 上）

Shap-E のインストール（Windows 上）

Blender のインストール

Shap-E のインストール（Windows 上）

Shap-E の動作確認（Windows 上）

Shap-E を使う Python プログラムの実行（Windows 上）