'caffe2 - create own MNIST lmdb dataset from numpy array

I am currently trying to create a MNIST lmdb dataset from my own data saved as numpy arrays data (shape (50000,1,28,28)) and label (shape (50000,)).

Does someone know how to convert numpy arrays to a lmdb dataset suitable for a caffe2 model?

KR ulk

numpymnistlmdbcaffe2


Sources

This article follows the attribution requirements of Stack Overflow and is licensed under CC BY-SA 3.0.

Source: Stack Overflow

Solution Source

Related Questions

I m trying to convert the pixels which is painted black to another color but ı got an error l

I m trying to convert the pixels which is painted black to another color but ı got an error l

I m trying to convert the pixels which is painted black to another color but ı got an error l

NumPy fancy indexing with provided output array

A way to get the local minima and maxima of a vector

Quick way to access first element in Numpy array with arbitrary number of dimensions?

How can I map True/False to 1/0 in a Pandas DataFrame?

Use numpy.load on file within zipfile

python(or numpy) equivalent of match in R

Exception thrown in PyCharm debug mode but not in run mode (seaborn, matplotlib)

Can numpy bincount work with 2D arrays?

Scikit Learn SVC decision_function and predict

How to generate a matrix with circle of ones in numpy/scipy

Creating a Linear Gradient Mask using opencv or

Extract subarray between certain value in Python