- Generated by
1.9.1
|
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.
|
Classes | |
| class | RemoveExtraInfoStage |
| Strips extra info for subcallables to save memory. More... | |
| class | CacheBeforeYieldStage |
| Caches results in a list and then yields them. More... | |
| class | SkipIfDumpExistsStage |
| Check if the output file is already generated, skip the run if so. More... | |
| class | MultiProcessingSpawnStage |
| Multiprocessing support stage. More... | |
| class | MultiProcessingGatherStage |
| Multiprocessing support stage. More... | |
Namespaces | |
| zigzag.stages.run_opt_stages | |
Functions | |
| def | get_threadpool (int|None nb_threads_if_non_existent) |
| def | close_threadpool () |
| def | terminate_threadpool () |
| def | raise_exception (Exception e) |
Variables | |
| logger = logging.getLogger(__name__) | |
| threadpool = None | |
1.9.1