Adding Variable in Tensor Class

1

I wanted to add an ID variable in Tensor class (pytorch/torch/_tensor.py at main · pytorch/pytorch · GitHub). I overloaded an init function inside the class Tensor.

class Tensor(torch._C.TensorBase):
    def __init__(self, other):
        super().__init__()
        print("Initializing Tensor...")
        self.ts_id = random.randint(0, 3000)
        print(f"Tensor ID: {self.ts_id}")

I observed that only Parameter tensors are only invoking the init method. For intermediate tensors, they are not calling this init method. My question is do intermediate tensors use another Tensor class for initialization?