v0.2.0-Complex, Images, Code clean up

Main new features

• Full Native Complex Number and Complex tensor support.
• Opening images as Ndarr.
• Saving Ndarr as images in multiple formats.
• Activation functions for Ndarr<T: Float,R>.
• Scan operation.
• Num tensor operations: max, min, sum, etc.
• Activation functions.

Changelog

• Code clean up.
• Implement From instead of Into trait.
• Float trait for Trig operations.
• Bug fix in scalar operations.
• Remove unnecessary Traits.
• Complex module.
• de_slice function.
• Scan operations.
• Singed tensor operations.
• README examples.
• Implementing sum operation for Ndarr.
• Hermitian transpose.
• Opening images and Ndarr.
• Implementing saving as RGB image for Ndarr<u8,3>.
• Implementing saving as RGB image for Ndarr<f32,3>.
• Implementing saving as Luma image for Ndarr<f32,2>.
• Implementing saving as Luma image for Ndarr<u8,2>.
• Activation functions: sigmoid, relu, leaky_relu, softmax

##Contributors
@S-Erik @JErnestoMtz

v0.3.0-Stable, Random, FFT

Main new features

• rapl is now in stable rust!!
• New Dim<R: Unsinged> struct.
• Activation functions.
• Random array creation.
• 1D and 2D FFT.
• linspace, logspace, geomspace implementation

Changelog

• Add approx method for testing operation with floating numbers.
• Matmul, Outer Product and Inner Product as methods of Ndarr.
• Add Random arrays.
• Add seeding for Random array creation.
• Normal and uniform Distribution.
• Numpy-like roll function.
• Indexing Ndarr with arrays.
• Code clean up in arithmetic operations.
• Fix co-broadcasting in scalar to tensor arithmetic operations.
• System independent paths in image module.
• Activation function.
• Improve display performance.
• Add display parameters.
• Reduce trait requirements in Ndarr.
• Add game example folder.
• Improve performance in slice

Contributors

@DeliciousHair @S-Erik @JErnestoMtz