Revert "Bump llama-cpp-python from 0.3.5 to 0.3.6"

[aponcedeleonch]

Reverts #688

Still getting the error when making make image-build

6okqk8/llama_cpp_python-0.3.6/vendor/llama.cpp/ggml/src/ggml-quants.c
23.66   ninja: build stopped: subcommand failed.
23.66
23.66
23.66   *** CMake build failed
23.66
23.66
23.66   at /usr/local/lib/python3.12/site-packages/poetry/installation/chef.py:164 in _prepare
23.66       160│
23.66       161│                 error = ChefBuildError("\n\n".join(message_parts))
23.66       162│
23.66       163│             if error is not None:
23.66     → 164│                 raise error from None
23.66       165│
23.66       166│             return path
23.66       167│
23.66       168│     def _prepare_sdist(self, archive: Path, destination: Path | None = None) -> Path:
23.66
23.66 Note: This error originates from the build backend, and is likely not a problem with poetry but with llama-cpp-python (0.3.6) not supporting PEP 517 builds. You can verify this by running 'pip wheel --no-cache-dir --use-pep517 "llama-cpp-python (==0.3.6)"'.
23.66
------
Dockerfile:20
--------------------
  19 |     # Configure Poetry and install dependencies
  20 | >>> RUN poetry config virtualenvs.create false && \
  21 | >>>     poetry install --no-dev
  22 |
--------------------
ERROR: failed to solve: process "/bin/sh -c poetry config virtualenvs.create false &&     poetry install --no-dev" did not complete successfully: exit code: 1

View build details: docker-desktop://dashboard/build/trusty-builder/trusty-builder0/86za5ao2rbxyv477b063gx6em
make: *** [image-build] Error 1

[yrobla]

Seems related to ggml-org/llama.cpp#10933