[WIP][Docs] Simple example of using MLX distributed

[stefpi]

related issue: #2930

proposed changes:

• add distributed layers into python/nn/layers for auto summarization
• add tensor parallelism inference example in docs
  • describe purpose and usage of ShardedToAllLinear and AllToShardedLinear layers
  • describe purpose and usage of Quantized versions
  • show simple example of combining layers together and its benefit
  • TP applied simple Llama model inference script
• add data parallelism training example in docs

Adding this as a draft PR as I work on these tasks. Please leave feedback on what I’ve done or have yet to do if you have any. Thanks!