Extendability refactors

[dakinggg]

This PR includes a few changes for increased extendability of the code:

• Fixes some typing
• Adds slice_attention_mask to MPTBlock
• Removes the need for extra imports in configuration_mpt.py just for HF checkpointing
• Factors out the attention config validation
• Makes the block class overridable on MPTModel
• Adds a registry for transforms that can be applied to the TrainConfig

Loss before and after:

[snarayan21]

few comments

[milocress]

Looks good pending GPU test fix

tests/models/layers/test_dmoe.py:135: in test_dmoe
    expert_parallel_group = device_mesh['expert_parallel'].get_group(0)
/usr/lib/python3/dist-packages/composer/trainer/_patch_pytorch.py:1041: in device_mesh__getitem__
    submesh = _mesh_resources.create_child_mesh(self, mesh_dim_names)
E   NameError: name '_mesh_resources' is not defined

[dakinggg]

@milocress the GPU test is unrelated. It will be fixed by the next composer release (which is why that test isn't marked as required yet)