Add docker model unload

[doringeman]

On top of #64.

You can launch docker/model-runner#46 using make docker-run.

$ MODEL_RUNNER_HOST=http://localhost:8080 docker model run ai/smollm2 hi
$ MODEL_RUNNER_HOST=http://localhost:8080 docker model unload --all
Unloaded 1 model(s).
$ # run again
$ MODEL_RUNNER_HOST=http://localhost:8080 docker model unload ai/smollm2
Unloaded 1 model(s).
$ # run again
$ MODEL_RUNNER_HOST=http://localhost:8080 docker model unload --backend llama.cpp ai/smollm2
Unloaded 1 model(s).

[crazy-max]

Same as #63 (comment)

[doringeman]

Fixed docs in #67.