Specifying NumPy Arrays with 2-Bit Dtype

Question

According to the docs, it's possible to specify different array dtypes:

dt = np.dtype('u1')   # 8-bit unsigned integer
dt = np.dtype('i4')   # 32-bit signed integer
dt = np.dtype('f8')   # 64-bit floating-point number
dt = np.dtype('c16')  # 128-bit complex floating-point number
dt = np.dtype('a25')  # 25-length zero-terminated bytes
dt = np.dtype('U25')  # 25-character string

However, the smallest unsigned integer dtype is 8-bit. Is there a way to create a 2-bit unsigned integer dtype?

Answer 1

According to the NumPy mailing list from November 2009, NumPy has 1-byte atomicity and so a 8-bit is the smallest unit. Even bool dtype uses a single byte.

Answer 2

I don't believe there is. Nowhere in the numpy docs, as far as i know, does it state that it is possible to define a 2 bit integer data type: https://www.numpy.org/devdocs/user/basics.types.html

Even a boolean data type which has a size of only 1 bit is stored as a byte in numpy.