Corresponding to some demos that we'd like to have done by mid march. This milestone should capture:

• Simple tests for all TTNN supported CCL ops in tt-mlir
• Simple tests that go e2e through tt-xla
• Simple tests that demonstrate "tensor" parallel and "batch" parallel
• Support for n300, LLM Box, TG, & ~BH in CI
• Manually Sharded Transformers
• Manually Sharded Bert
• Preliminary Shardy Support
• Investigation into PyTorch / file issue torch-mlir to start discussion
• Investigate automatic shard solver
• Simple training