fx quant: make load_arg function more clear (#43923)
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/43923
Readability improvements to `Quantizer.convert.load_arg`, makes
things easier to read.
1. add docblock
2. `arg` -> `arg_or_args`, to match what's actually happening
3. `loaded_arg` -> `loaded_args`, to match what's actually happening
Test Plan:
```
python test/test_quantization.py TestQuantizeFxOps
python test/test_quantization.py TestQuantizeFx
```
Imported from OSS
Reviewed By: jerryzh168
Differential Revision: D23438745
fbshipit-source-id: f886b324d2e2e33458b72381499e37dccfc3bd30