fix(backend): count the S3TokenizerV2 repo in the Chatterbox MLX cache check; register the backend with PyInstaller

Review follow-ups: mlx-audio's Model.from_pretrained fetches the S3
speech tokenizer from mlx-community/S3TokenizerV2 (~470 MB) separately
from the chatterbox checkout, so _is_model_cached now requires both
repos (same shape as the Hume backend's codec check) and the config's
size_mb reflects the real footprint. backend.backends.chatterbox_mlx_backend
is a function-level import that PyInstaller's graph will not see, so it
is added to the Apple Silicon hidden-import list in build_binary.py and
voicebox-server.spec next to mlx_backend.

backend/build_binary.py

@@ -442,6 +442,8 @@ def build_server(cuda=False, rocm=False):
                 "mlx_lm",
                 "--hidden-import",
                 "backend.backends.qwen_llm_backend",
+                "--hidden-import",
+                "backend.backends.chatterbox_mlx_backend",
                 "--collect-submodules",
                 "mlx",
                 "--collect-submodules",