[Quant][FX] lower functional conv_transpose ops (#97126)
**Summary**
Support quantizing and lowering functional `conv_transpose1d`, `conv_transpose2d` and `conv_transpose3d`.
Please note that
- `conv_tranpose + relu` fusion is not supported. Remember to keep `relu` node in graph when lowering.
- `conv_tranpose` requires `per-tensor` scheme for weight. Use default `qconfig_mappings` instead of deprecated `qconfig_dict` for test cases.
**Test plan**
python test/test_quantization.py -k test_conv_transpose_not_reference
python test/test_quantization.py -k test_conv_transpose_reference
python test/test_quantization.py -k test_conv_transpose_relu_not_reference
python test/test_quantization.py -k test_conv_transpose_relu_reference
Pull Request resolved: https://github.com/pytorch/pytorch/pull/97126
Approved by: https://github.com/jgong5, https://github.com/jerryzh168
