Support multi-gpu transformers inference

[sunyuhan19981208]

Add a script for multi-gpu inference

[airaria]

Is it possible to use one script for both single-gpu and multi-gpu inference (for instance, users can launch the script with --single-gpu (default) or --multi-gpu) ? so that we only have to maintain one script.

[ymcui]

Thank you for your contributions. We are indeed interested in your multi-gpu implementation. As requested, please consider merging the original inference script and yours into a unified one, so that the users can use with more flexibilities.

[sunyuhan19981208]

Thank you for your contributions. We are indeed interested in your multi-gpu implementation. As requested, please consider merging the original inference script and yours into a unified one, so that the users can use with more flexibilities.

Okay, I will finish that work, thank you for your concern.

[sunyuhan19981208]

Is it possible to use one script for both single-gpu and multi-gpu inference (for instance, users can launch the script with --single-gpu (default) or --multi-gpu) ? so that we only have to maintain one script.

Thank you for your advice. I will complete the task as per your instructions.

[airaria]

have tested the code and it looks ok