Align `_choose_scale_float8`'s use of `block_size` with other code

[jerryzh168]

Recently I found that for _choose_scale_float8:

len(block_size) == 0 for per tensor:

ao/torchao/quantization/quant_primitives.py

Line 2303 in b4ec4cb

if len(block_size) == 0:

keepdim=True for others:

ao/torchao/quantization/quant_primitives.py

Line 2314 in b4ec4cb

max_abs = tensor_reshaped.abs().amax(dim=reduction_dims, keepdim=True)

While for _choose_qparams_affine (for int):

we use keepdim=False:

ao/torchao/quantization/quant_primitives.py

Lines 1552 to 1553 in b4ec4cb

	min_val = torch.amin(input, dim=reduction_dims, keepdim=False)
	max_val = torch.amax(input, dim=reduction_dims, keepdim=False)

that will make per tensor quantization to have a scalar scale automatically.

I think we should fix this discrepancy to reduce confusion, by aligning _choose_scale_float8 with the int one, i.e. change keepdim=True to keepdim=False and make sure the tests work.

we can also add some docs afterwards to both ops to clarify.