custom gpu option

[Dion11235]

This change resolves the issue of loading the pre-trained model into any other gpu for systems with multiple multiple gpu.
setting visible device is a bypass but I needed to use two gpus in parallel for modelling and reaction mapping in a common environment, so this feature seemed useful. Thanks. Please let me know about your thoughts and apologies if anything is done inappropriately as this is my first PR.
signed-off by : Dipan Banik [email protected]

[cla-bot]

Thank you for your pull request and welcome to our community. We could not parse the GitHub identity of the following contributors: dipan.banik.
This is most likely caused by a git client misconfiguration; please make sure to:

1. check if your git client is configured with an email to sign commits git config --list | grep email
2. If not, set it up using git config --global user.email [email protected]
3. Make sure that the git commit email is configured in your GitHub account settings, see https://github.com/settings/emails

[cla-bot]

Thank you for your pull request and welcome to our community. We could not parse the GitHub identity of the following contributors: dipan.banik.
This is most likely caused by a git client misconfiguration; please make sure to:

1. check if your git client is configured with an email to sign commits git config --list | grep email
2. If not, set it up using git config --global user.email [email protected]
3. Make sure that the git commit email is configured in your GitHub account settings, see https://github.com/settings/emails