v-0.8.3

specify dtype in linspace and arange function default dtype would be a float

v-0.8.2

New

• .hard_sigmoid(symbol: Optional[str]=None)
• .hard_shrink(symbol: Optional[str]=None)
• .hard_tanh(symbol: Optional[str]=None)
• .hard_swish(symbol: Optional[str]=None)
• .leaky_relu(negative_slope: Union[int,float]=0.01, inplace:bool=False, symbol: Optional[str]=None)

v-0.8.1

v0.8.1

v-0.8.1

v-0.8.0

A sub-module has been added called extra

• exp2(matrix: Matrix, symbol: Optional[str]=None)
• exp3(matrix: Matrix, symbol: Optional[str]=None)
• expn(matrix: Matrix, symbol: Optional[str]=None)
• expit(matrix: Matrix, symbol: Optional[str]=None)
• expm1(matrix: Matrix, symbol: Optional[str]=None)
• entr(matrix: Matrix, symbol: Optional[str]=None)
• i0(matrix: Matrix, ,terms: int=50 symbol: Optional[str]=None)
• i0e(matrix: Matrix, terms: int=50, symbol: Optional[str]=None)
• i1(matrix: Matrix, terms: int=50, symbol: Optional[str]=None)
• i1e(matrix: Matrix, terms: int=50, symbol: Optional[str]=None)
• logit(matrix: Matrix, eps=None, symbol: Optional[str]=None)

v-0.7.5

v0.7.5

v-0.7.5

v-0.7.4

changes:

• rand(shape:Tuple[int,int], dtype:Type, seed:None|int=None, symbol:Optional[str]=None) and .rand_like( ... ) now supports different dtypes, e.g int, float and complex
• .to_list(keepdim:bool=True) now can flatten the matrix into 1D array or can return 2D array list

updates:

• .squeeze() returns the 1D array in the form of python list
• .real and .imag both are the matrix properties that returns the real and imaginary parts of a matrix

v-0.7.3

changes:

1. by default the datatype for matrix would a float, for complex numbers you need to specify dtype=complex
2. .mat_pow(n: int) deprecated, instead use .matpow(n: int)
3. .concatenate( ... ) deprecated, instead use .concat( ... ) parameters are the same as before
4. major changes in matrix representation see README.

added:

1. inverse hyperbolic trigonometric ratios were added
   • .asinh()
   • .acosh()
   • .atanh()
   • .acosech()
   • .asech()
   • .acoth()
2. multiply matrices with vectors, calling a method called .vecmul(vector: List[Union[int,float,bool,complex]])

v-0.7.2

v0.7.2

v-0.7.2

v-0.7.1

some bugs and issues were resolved

v-0.7.0

Now Matrix supports complex numbers 👍