Generation of Acoustic Wave Dataset in vectorial form¶

Libraries import¶

In [4]:
from pyawd import VectorAcousticWaveDataset3D, Marmousi
import torch
import matplotlib.pyplot as plt

GENERATING = True

Dataset generation¶

In [2]:
size = 100
nx = 32
ddt = 0.01
dt = 2
t = 10
batch_size = 10

if GENERATING:
    dataset = VectorAcousticWaveDataset3D(size=size, nx=nx, ddt=ddt, dt=dt, t=t, 
                                          interrogators=[(-nx//4, 0, nx//2-1), 
                                                         (nx//4, 0, nx//2-1)])
    torch.save(dataset, '../../../local/dev.pt')
else:
    dataset = torch.load('../../../local/dev.pt')
dataloader = torch.utils.data.DataLoader(dataset, batch_size=batch_size, shuffle=True)

We plot an example of interrogators measurement through time:

In [6]:
dataset.plot_interrogators_response(0)
No description has been provided for this image

We show an example of visual representation:

In [4]:
dataset.generate_video(0, "dataset_generation", 100)

  0%|          | 0/100 [00:00<?, ?it/s]
In [5]:
%%HTML
<video width="640" controls>
  <source src="dataset_generation.mp4" type="video/mp4">
</video>
In [ ]: