conscell

This implementation is numerically unstable. For large x e.g. x=1000 it will cause the following error: OverflowError: math range error

@tf318 This implementation fails tests: FAILED test_engine.py::test_sanity_check - assert 128.0 == 46.0 FAILED test_engine.py::test_more_ops - assert 15.166180758017475 < 1e-06

@Weikang01 If self.data is negative and other is a `Value`, this implementation will cause a `ValueError: math domain error`. Division is implemented as follows: ``` def __truediv__(self, other): # self...

@Weikang01 In negative case perhaps it makes sense to set gradient to `math.nan`, something like this: ``` >>> import torch >>> a = torch.tensor(-2.0, requires_grad=True) >>> b = torch.tensor(-3.0, requires_grad=True)...

By default this port is intended for use with Jack Palevich's Android Terminal Emulator and must be placed here: '/data/data/jackpal.androidterm/app_HOME'. If you want to place it somewhere else, you can...

> but the compilation went wrong Could you please provide more details about compilation issues. > Hugs> :l Data.Graph > ERROR "/data/data/jackpal.androidterm/app_HOME/hugs/lib/hugs/packages/hugsbase/Hugs/ST.hs":51 - Syntax error in type expression (unexpected `.')...

This implementation is numerically unstable. For large x e.g. x=1000 it will cause the following error: OverflowError: math range error

@gordicaleksa Could you please provide an example when the computational graph is not a DAG?

This won't cause an infinite recursion, just unnecessary __radd__ and __rmul__ calls. Also, it is a duplicate of https://github.com/karpathy/micrograd/pull/39.

No, because this would cause an error: ``` >>> from micrograd import engine >>> v = engine.Value(3) >>> v + 3 Value(data=6, grad=0) >>> 3 + v Traceback (most recent...