/Distributed_Deep_Learning_Inference

The Python implementation of distributed auction algorithm for multiple assignment (DAMA) for CNN inference.

Primary LanguageJupyter Notebook

Distributed_Deep_Learning_Inference

The Python implementation of distributed auction algorithm for multiple assignment (DAMA) of CNN inference.

Requirements:

  • Torch 1.3.0
  • Tensorflow-gpu 1.12.0

Start Server:

python grpc_server.py 50051

python grpc_server.py 50052

Start Client:

python grpc_client.py

Under Development - Issues Remaining to Solve

  • Have a generalized pytorch config file that supports all types of CNNs
  • Fix the process launch and termination for each server (Multiprocessing)
  • Refactor the code and use single scripts for launching servers using a common config file