Can I pause Experiment Manager App and then continue to run experiments later without losing progress?

I want to run a bayesian optimization experiment which could lead to running of 30 or so networks. Can I pause the experiment manager when I need to use matlab or when I need to close my computer etc... and then continue without losing any progress? Or can I train the networks in batches for ex can I train 2 and then perhaps on another session another 2. Thanks

Answer 1

[Milan Bansal]

Hi,

As per my understanding, you want to pause the Experiment Manager App and save the results of the completed experiments such that no progress is lost and continue working later. Also you want to know if is possible to train the network in the batches.

There are two options available in Experiment Manager App.

1. Stop : It stops the running trial and saves its results. It is possible to export the training results in this case.
2. Cancel : It cancels the running trial and discards the results.

Refer to the documentation link given below to know more about "Stop" and "Cancel" options in Experiment Manager App.

https://www.mathworks.com/help/deeplearning/ref/experimentmanager-app.html?searchHighlight=experiment%20manager%20app&s_tid=srchtitle_support_results_1_experiment%20manager%20app#mw_b4f8fdb1-d24b-4afc-90e2-959beca77a29 

Moreover, if Parallel Computing Toolbox is available, it is possible to run multiple trials at the same time or run a single trial on multiple GPUs in a cloud. Refer to the link to learn more.

https://www.mathworks.com/help/deeplearning/ug/use-experiment-manager-to-train-networks-in-parallel.html 

If MATLAB Parallel Server is available, experiments can be offloaded as batch jobs in remote server, so that it is possible to use MATLAB and even close the current MATLAB session while training. Refer to the link to learn more.

https://www.mathworks.com/help/deeplearning/ug/offload-experiment-as-batch-job-to-cluster.html