- Use ptr.contents, not ptr in `np.array`
- Get dtype from return type on annotated signature
- Explicitly set copy=True and dtype on np.array
  - Should not strictly be necessary since pointer is typed

Catch StateReloadError and add logging if runs into this when running.

- Fix loading state (from_buffer -> from_buffer_copy since bytes aren't mutable)
- Add tests (E2E, errors when should, reloads successfully, logits correct, etc.)

Have to set LLAMA_TEST_MODEL to point to model path in order to get this
to run.

D'oh

- Check when saving that model doesn't need logits
- Ad note in `reload_from_cache` state to revisit

- Make default to *not* save logits
  - Error if needed and save_logits False in build_cache
-  Handle reloading with/without scores if needed + available

- Add more tests
- Make llama_state / small_model module scope (so don't need to reload for each test)
  - Setting env var in `.env` file