[5597780] Add support for FP16-only custom ops#460

gcunhase merged 3 commits intoNVIDIA:mainNVIDIA/TensorRT-Model-Optimizer:mainfrom

gcunhase:dev/gcunhasergio/autocast_custom_op_add_fp16_only_support_5597780gcunhase/TensorRT-Model-Optimizer:dev/gcunhasergio/autocast_custom_op_add_fp16_only_support_5597780Copy head branch name to clipboard