Created
May 28, 2021 05:10
-
-
Save 1pha/657d6f06c7acc6414075667552c96b3a to your computer and use it in GitHub Desktop.
Understanding collate_fn
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
import torch | |
from torch.utils.data import Dataset, DataLoader | |
import numpy as np | |
class MyDataset(Dataset): | |
def __init__(self): | |
x = np.random.randint(10, size=[1000, 3]) # 1000 3-dim samples | |
self.x = [x[i].tolist() for i in range(1000)] | |
y = np.random.randint(low=0, high=2, size=(1000,)) | |
self.y = [y[i] for i in range(1000)] | |
def __len__(self): | |
return len(self.x) | |
def __getitem__(self, idx): | |
return self.x[idx], self.y[idx] | |
def collate_fn(batch): | |
print(f"batch: {batch}") | |
data_list, label_list = [], [] | |
for _data, _label in batch: | |
data_list.append(_data) | |
label_list.append(_label) | |
return torch.Tensor(data_list), torch.LongTensor(label_list) | |
if __name__ == "__main__": | |
dataset = MyDataset() | |
print(len(dataset)) | |
print(dataset[-1]) | |
print(f"\nWITHOUT COLLATE_FN") | |
dataloader = DataLoader(dataset, batch_size=3, shuffle=False) | |
for data, label in dataloader: | |
print(data) | |
print(label) | |
break | |
print(f"\nWITH COLLATE_FN") | |
dataloader = DataLoader(dataset, batch_size=3, shuffle=False, collate_fn=collate_fn) | |
for data, label in dataloader: | |
print(data) | |
print(label) | |
break |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Thanks to: https://gist.github.com/subhadarship/e5a60bd3ef7ef845348325bfb4d9ddc1/ 🙏