Fixed f-string printing of NF4Tensors

[awgu]

import torch
from torchao.dtypes import NF4Tensor

weight = torch.randn((16, 16), device="cuda", dtype=torch.bfloat16)
block_size, scaler_block_size = 2, 2
nf4_weight = NF4Tensor.from_tensor(weight, block_size, scaler_block_size)
print(f"{nf4_weight}")

File "/data/users/andgu/pytorch/test_nf4.py", line 13, in main
    print(f"{nf4_weight}")
  File "/data/users/andgu/pytorch/torch/_tensor.py", line 986, in __format__
    return handle_torch_function(Tensor.__format__, (self,), self, format_spec)
  File "/data/users/andgu/pytorch/torch/overrides.py", line 1646, in handle_torch_function
    result = torch_func_method(public_api, types, args, kwargs)
  File "/data/users/andgu/ao/torchao/dtypes/nf4tensor.py", line 797, in __torch_function__
    return func(*args, **kwargs)
  File "/data/users/andgu/pytorch/torch/_tensor.py", line 989, in __format__
    return object.__format__(self, format_spec)
  File "/data/users/andgu/ao/torchao/dtypes/nf4tensor.py", line 754, in __str__
    return self.to(torch.float32).__str__()
  File "/data/users/andgu/ao/torchao/dtypes/nf4tensor.py", line 778, in __torch_dispatch__
    return NF4_OPS_TABLE[func](func, args, kwargs)
  File "/data/users/andgu/ao/torchao/dtypes/nf4tensor.py", line 277, in _to_copy
    return args[0][0].get_original_weight().to(args[1]["dtype"]).to(args[1]["device"])
KeyError: 'device'

Note that args[1] only has a "dtype" key, not "device":

args[1]: {'dtype': torch.float32}

[pytorch-bot]

🔗 Helpful Links

🧪 See artifacts and rendered test results at hud.pytorch.org/pr/pytorch/ao/297

• 📄 Preview Python docs built from this PR

Note: Links to docs will display an error until the docs builds have been completed.

❗ 1 Active SEVs

There are 1 currently active SEVs. If your PR is affected, please view them below:

• Upgrade MacOS runner to 14

✅ No Failures

As of commit 613d577 with merge base 38dad9b ():
💚 Looks good so far! There are no failures yet. 💚

This comment was automatically generated by Dr. CI and updates every 15 minutes.