Add PyTorch examples into CI

BladeDISC is heavily developing. The updates of TorchMLIR, PyTorch, and MHLO/XLA/TensorFlow versions and changes in BladeDISC might introduce unstable features that break end-to-end optimizations of examples.

Some examples can not pass the CI check, we need to push some prs to fix these bugs and add it to the CI system:

- [ ] [AlphaFold](https://github.com/alibaba/BladeDISC/tree/main/examples/PyTorch/Inference/CUDA/AlphaFold)
- [x] [BERT](https://github.com/alibaba/BladeDISC/tree/main/examples/PyTorch/Inference/CUDA/BERT) @tanyokwok
- [x] [ResNet](https://github.com/alibaba/BladeDISC/tree/main/examples/PyTorch/Inference/CUDA/ResNet) @tanyokwok
- [x] [S2T](https://github.com/alibaba/BladeDISC/tree/main/examples/PyTorch/Inference/CUDA/S2T) @Yancey1989
- [x] [T5](https://github.com/alibaba/BladeDISC/tree/main/examples/PyTorch/Inference/CUDA/T5)
- [ ] [CPU/albert](https://github.com/alibaba/BladeDISC/tree/main/examples/PyTorch/Inference/CPU/albert)
- [x] trigger examples CI daily @zzpmiracle

Start with the PyTorch example paths for AlphaFold and CPU/albert, then inspect how the checked BERT, ResNet, S2T, and T5 examples are triggered in CI. Reproduce the failing checks and address the example-specific issues; done means the remaining examples pass end-to-end optimization checks and are included in the CI system.

Written by the indexing model from the issue text.