This repository provides a Python implementation for:
Jiaming Qiu, Ruiqi Wang, Ayan Chakrabarti, Roch Guérin, and Chenyang Lu, "Adaptive Edge Offloading for Image Classification Under Rate Limit" [paper] [slides] [video]
This work is an extension of "Real-time Edge Classification: Optimal Offloading under Token Bucket Constraints" [paper] [code]
We recommend a recent Python 3.7+ distribution of Anaconda with numpy, tensorflow 2, and matplotlib installed.
To reproduce the results presented in the paper, you may download and extract our pre-computed data merged from multiple simulations. The files should be extracted into the viz/data/ sub-directory, which allows you to directly run the note-books in the visualization step.
You will need to first download an npz file containing the pre-computed offloading metrics for the ILSVRC validation set with the weak and strong classifier pair we reported in the paper (VGG-style 16-layer vs. OFA 595MFlops). You can put the file in the data/ sub-directory. Check our previous work for more details on computing offloading metrics with your own dataset and classifier pair.
train.py in the root directory is the main script for training the DQN model. Here is a typical example run:
# Start running 3k Q-value iterations with a learning rate of 10^-3.
./train.py data/ofa_top5.npz wts --tr 0.25 --tb 2 --stype 1 --itype 1 --lr 3 --maxiter 3000
# Reward and loss have saturated. Drop the learning rate to 10^-4 and run another 1k iterations.
./train.py data/ofa_top5.npz wts --tr 0.25 --tb 2 --stype 1 --itype 1 --lr 4 --maxiter 4000Running the script with -h gives you a detailed list of parameters.
After you trained the DQN model, weights are stored in the specified directory (wts/ in the above example). The model is then ready for simulation on the test sequences with simulate.py. Make sure you point the script to the right directory to load the model and use consistent token bucket and sequence configurations throughout training and simulation. For example, to run simulation with the model trained above, you should use:
# Set the model directory to wts/ and use the same configuration.
./simulate.py data/ofa_top5.npz wts save --tr 0.25 --tb 2 --stype 1 --itype 1The simulation results (Average loss of the DQN, MDP, Baseline and Lower Bound policies, as well as that of the weak and strong classifiers) are stored in the specified directory (save/ in the above example). You can compare the performance of different policies, or the performance of the same policy accorss different configurations by training DQN models and running simulations accorss these configurations.
The actual library for loading data, generating sequences, computing policies and simulations is in the lib directory. Please take a look at its README.md file for documentation.
We provide separate jupyter notebooks in viz/notebook/ to visualize (either downloaded or generated) results, producing the figures included in the paper.
This code is being released under the MIT License. Note that the offloading metrics are computed following the procedure developed in our previous work, and the OFA results files were generated from the models and code provided by its authors.
This work was partially supported by the National Science Foundation under grants no. CPS-1646579 and CNS-2006530, and the Fullgraf Foundation. Any opinions, findings, and conclusions or recommendations expressed in this material are those of the authors, and do not necessarily reflect the views of the National Science Foundation and the Fullgraf Foundation.
Ayan Chakrabarti extensively contributed to this implementation by collaborating with Jiaming Qiu.