v0.0.4

Releases 0.0.4

I am trying to make it platform independent.

Bug Solution

In this release, I fixed the problem in matrix multiplication.

Feature addition

• add anygrad.FloatTensor, anygrad.IntTensor, and anygrad.BoolTensor.
• add the version information access like this import anygrad; anygrad.__version__
• add dtypes like int32, int64 and bool
• add anygrad.cast function to convert the one tensor type to another.
  and many more ...

Full Changelog: 0.0.3...0.0.4