Blank Audio getting generated in case of using TRT optimization

[tool1996-oss]

Checks

• This template is only for bug reports, usage problems go with 'Help Wanted'.
• I have thoroughly reviewed the project documentation but couldn't find information to solve my problem.
• I have searched for existing issues, including closed ones, and couldn't find a solution.
• I am using English to submit this issue to facilitate community communication.

Environment Details

1. I am using 4xA5000 GPUs on linux 20.04

Steps to Reproduce

1. I am attaching my custom file link
2. I used the bencmark.py file with my custom checkpoint is order to generate audio but its completly empty empty audio
3. The above problem can be easily solved by setting remove_input_padding as False although dont know why is that happening.
4. After converting the model in fp32 (previously it was fp16) the model is producing gibberish audio

✔️ Expected Behavior

The model should be able to generate properly

❌ Actual Behavior

I printed the generated audio it is giving only nan values
tensor([[[nan, nan, nan, ..., nan, nan, nan], [nan, nan, nan, ..., nan, nan, nan], [nan, nan, nan, ..., nan, nan, nan], ..., [nan, nan, nan, ..., nan, nan, nan], [nan, nan, nan, ..., nan, nan, nan], [nan, nan, nan, ..., nan, nan, nan]]], device='cuda:0')

[JohnHerry]

The same issure. what is the progress?