Transformer Engine prebuilt Wheel

## Official Release, prebuilt wheel
```
260121_a/transformer_engine-2.5.0+f05f12c9-cp312-cp312-linux_x86_64.whl
  * build environment: nvcr.io/nvidia/pytorch:25.06-py3
    * torch: 2.8.0a0+5228986c39.nv25.06
    * cuda : 12.9
    * cudnn: 91002 (9.10.2)
    * TE branch release_v2.11 c188b5 dec 16 (CUDA_ARCHS="90")
```

```
250823_a/transformer_engine-2.5.0+f05f12c9-cp312-cp312-linux_x86_64.whl
  * build environment: nvcr.io/nvidia/pytorch:25.06-py3
    * torch: 2.8.0a0+5228986c39.nv25.06
    * cuda : 12.9
    * cudnn: 91002 (9.10.2)
    * TE stable v2.5 (CUDA_ARCHS="90;100;120")
```

## Unofficial, prebuilt wheel for local experimentation
```
unofficial/251115_a/transformer_engine-2.5.0+c2056e01-cp312-cp312-linux_x86_64.whl
  * build environment: nvcr.io/nvidia/nemo:25.07
    torch: 2.8.0a0+5228986c39.nv25.05
    cuda : 12.9
    cudnn: 91100
    * TE vuiseng9's fork:v2.5+nv4f_mx8b (CUDA_ARCHS=100)
```

```
unofficial/251025_a/transformer_engine-2.8.0+7cde9a33-cp312-cp312-linux_x86_64.whl
  * build environment: nvcr.io/nvidia/nemo:25.09
    torch: 2.8.0a0+5228986c39.nv25.06
    cuda : 12.9
    cudnn: 91100 (9.11.0)
    * TE vuiseng9's fork:v2.8 ++ (CUDA_ARCHS=100a)
```
```
unofficial/251020_a/transformer_engine-2.5.0+c2056e01-cp312-cp312-linux_x86_64.whl
  * build environment: nvcr.io/nvidia/nemo:25.07
    * torch: 2.8.0a0+5228986c39.nv25.05
    * cuda : 12.9
    * cudnn: 91100 (9.11.0)
    * TE: vuiseng9's fork: branch v2.5+nvfp4 (c2056e01)
    # CUDA_ARCHS="100"
```
```
unofficial/251015_a/transformer_engine-2.5.0+010e0427-cp312-cp312-linux_x86_64.whl
  * build environment: nvcr.io/nvidia/pytorch:25.06-py3
    * torch: 2.8.0a0+5228986c39.nv25.06
    * cuda : 12.9
    * cudnn: 91002 (9.10.2)
    * TE: vuiseng9's fork: branch v2.5+nvfp4 (010e0427)
    # CUDA_ARCHS="100"
```