[Benchmark] SGD on Logistic Regression Benchmark Completes

cornell-zhang / heterocl

HeteroCL: A Multi-Paradigm Programming Infrastructure for Software-Defined Heterogeneous Computing

https://cornell-zhang.github.io/heterocl/

Apache License 2.0

322 stars 92 forks source link

[Benchmark] SGD on Logistic Regression Benchmark Completes #47

Closed seanlatias closed 6 years ago

seanlatias commented 6 years ago

In this pull request, I also added several APIs for NumPy arrays. First, there are "pack" and "unpack" API that can pack and unpack a NumPy array to another NumPy array given the input and output data type. Second, you can cast a NumPy array to another one by specifying a HeteroCL data type. Thus, fixed-point casting is also viable.

Also fixed other minor bugs in GetSlice and SetSlice operations.