* moved most FAISS_ASSERT calls to C++ exceptions, and adjusted
  memory allocation to avoid mem leaks

* added an IndexIVFScalarQuantizer type that offers an
  intermediate compression between IVFFlat and IVFPQ

* support removal of indices in IndexIDMap / IndexFlat combination

* various fixes in GPU code

  memory allocation to avoid mem leaks

* added an IndexIVFScalarQuantizer type that offers an
  intermediate compression between IVFFlat and IVFPQ

* support removal of indices in IndexIDMap / IndexFlat combination

* various fixes in GPU code

- better selection of training sets for PQ and preprocessing
- GPU parameter object
- IndexIDMap fixed
- fixed redo bug in clustering