[MPS] Fix MPS clamp issue with different dtypes between input and min/max tensors (#105747)
- Fix the FP16 clamp issue (FP32 and FP16 are not broadcast compatible)
- Fix clamp (cached graph nodes were previously replaced with the cast version)
Pull Request resolved: https://github.com/pytorch/pytorch/pull/105747
Approved by: https://github.com/kulinseth