Support packaging sharded models

[ekcasey]

Bumps model-distribution to include docker/model-distribution#123. This allows packaging of sharded models.

Using DMR build from docker/model-runner#133:

## Package
> docker model package --gguf /Users/emilycasey/Library/Caches/llama.cpp/Felladrin_gguf-sharded-Q4_K_S-DeepSeek-R1-Distill-Qwen-1.5B_model.shard-00001-of-00006.gguf shard/test
Adding GGUF file from "/Users/emilycasey/Library/Caches/llama.cpp/Felladrin_gguf-sharded-Q4_K_S-DeepSeek-R1-Distill-Qwen-1.5B_model.shard-00001-of-00006.gguf"
Loading model to Model Runner...
Transferred: 97.52 MB
Model loaded successfully

## Run
> docker model run shard/test
Interactive chat mode started. Type '/bye' to exit.
> hi
Hello! How can I assist you today?
> /bye
Chat session ended.

## Inspect bundle
> tree ~/.docker/models/bundles
/Users/emilycasey/.docker/models/bundles
└── sha256
    ├── 608d88953bbbb3d41964dd205882a0081227a50378050ee20a84994fe9ed005a
    │   ├── config.json
    │   ├── model-00001-of-00006.gguf
    │   ├── model-00002-of-00006.gguf
    │   ├── model-00003-of-00006.gguf
    │   ├── model-00004-of-00006.gguf
    │   ├── model-00005-of-00006.gguf
    │   └── model-00006-of-00006.gguf

[xenoscopic]

It might be helpful to avoid confusion if package's Short description included something like "Sharded GGUF files can be package by specifying the path to the first GGUF shard" (or something similar, I'm not sure the exact way to say it).

[ekcasey]

It might be helpful to avoid confusion if package's Short description included something like "Sharded GGUF files can be package by specifying the path to the first GGUF shard" (or something similar, I'm not sure the exact way to say it)

I couldn't find a wording that was short enough to look okay in docker model --help output so I added a Long help text with more details to be shown only on docker model package --help.