How could I deal with this issue when I`m running main.py?

JedMills / MTFL-For-Personalised-DNNs

Code for 'Multi-Task Federated Learning for Personalised Deep Neural Networks in Edge Computing', published in IEEE TPDS.

86 stars 20 forks source link

How could I deal with this issue when I`m running main.py? #5

Closed JacksonVation closed 2 years ago

JacksonVation commented 2 years ago

JedMills commented 2 years ago

Hi CoolYesWow,

I've updated the README with more detailed instructions on running the simulations.

Warm regards, Jed

On Mon, Jul 18, 2022 at 5:53 PM CoolYesWow @.***> wrote:

[image: image] https://user-images.githubusercontent.com/84269313/179562681-bc4b62bf-c463-4f4c-84a1-6019230e23e9.png

— Reply to this email directly, view it on GitHub https://github.com/JedMills/MTFL-For-Personalised-DNNs/issues/5, or unsubscribe https://github.com/notifications/unsubscribe-auth/AFLNUCMDY5RAT7FREP4NI43VUWDW3ANCNFSM5345I6BQ . You are receiving this because you are subscribed to this thread.Message ID: @.***>

JacksonVation commented 2 years ago

Hi Jed,

Its great to hear from you. And this really helps me a lot when Im stuck in training models! And now Im starting it~~ Its pretty cool. So when I`m done, how could I evaluate the performance of these models while I do not have any Raspberry Pi?

Warmly, Cool

JedMills commented 2 years ago

Hi Cool,

I'm glad it helps! I've also just added the file "Fig-5-settings.md" containing all the hyperparameters used for the experiments in Fig 5, which another user had requested.

The experiments from Fig. 4, Fig. 5, Table 2 and Table 3 are all simulated (run on a workstation), so if you want to reproduce those you will not need an RPi. Only the results from Table 4 (where we recorded the time taken to run each round of the algorithms) were performed on RPi's.

Warm regards, Jed