`thunder.jit`ted `Tensor.masked_fill` and `Tensor.masked_fill_` return `torcdh.float32` tensor even when input is `torch.int64`

[crcrpar]

Note: If you have a model or program that is not supported yet but should be, please use the program coverage template.

🐛 Bug

As per title.

To Reproduce

import torch
import thunder

if __name__ == "__main__":
    with torch.device("cuda"):
        a = torch.randint(0, 64, size=(384, 16), dtype=torch.int64)
        b = a.clone().detach()
        mask = torch.randn(size=(384, 16)).to(torch.bool)

    def f(a, mask):
        return a.masked_fill(mask, 0.0)

    expected = f(a, mask)
    jitted = thunder.jit(f)
    actual = jitted(a, mask)

    print(f"# `Tensor.masked_fill`: {expected.dtype = }, {actual.dtype = }")

$ python a.py
# `Tensor.masked_fill`: expected.dtype = torch.int64, actual.dtype = torch.float32

Expected behavior

Returns the same dtype as the pytorch result.

Environment

• PyTorch Version (e.g., 1.0):
• OS (e.g., Linux):
• How you installed PyTorch (conda, pip, source):
• Build command you used (if compiling from source):
• Python version:
• CUDA/cuDNN version:
• GPU models and configuration:
• Any other relevant information:

Additional context