ZigZag - Deep Learning Hardware Design Space Exploration
This repository presents the novel version of our tried-and-tested hardware Architecture-Mapping Design Space Exploration (DSE) Framework for Deep Learning (DL) accelerators. ZigZag bridges the gap between algorithmic DL decisions and their acceleration cost on specialized accelerators through a fast and accurate hardware cost estimation.
zigzag.utils Namespace Reference

Classes

class  UniqueMessageFilter
 Prevents the logger from filtering duplicate messages. More...
 
class  DiGraphWrapper
 Wraps the DiGraph class with type annotations for the nodes. More...
 

Functions

int hash_sha512 (Any data)
 Hashes the input data using SHA-512. More...
 
Any pickle_deepcopy (Any to_copy)
 
def pickle_save (str to_save, str path)
 
def pickle_load (str path)
 
dict[str, Any]|list[dict[str, Any]] open_yaml (str path)
 
Any json_repr_handler (Any obj, bool simple=False)
 Recursively converts objects into a json representation. More...
 

Variables

 T = TypeVar("T")
 

Function Documentation

◆ hash_sha512()

int zigzag.utils.hash_sha512 ( Any  data)

Hashes the input data using SHA-512.

Here is the caller graph for this function:

◆ json_repr_handler()

Any zigzag.utils.json_repr_handler ( Any  obj,
bool   simple = False 
)

Recursively converts objects into a json representation.

Here is the caller graph for this function:

◆ open_yaml()

dict[str, Any] | list[dict[str, Any]] zigzag.utils.open_yaml ( str  path)
Here is the caller graph for this function:

◆ pickle_deepcopy()

Any zigzag.utils.pickle_deepcopy ( Any  to_copy)
Here is the caller graph for this function:

◆ pickle_load()

def zigzag.utils.pickle_load ( str  path)

◆ pickle_save()

def zigzag.utils.pickle_save ( str  to_save,
str  path 
)

Variable Documentation

◆ T

T = TypeVar("T")