[quant] Add testing coverage for 4-bit embedding_bag sparse lookup op (#47328)
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/47328
Extend tests to cover case for pruned weights with mapping table.
Support for 8-bits sparse lookup to follow
Test Plan:
python test/test_quantization.py TestQuantizedEmbeddingOps
Imported from OSS
Reviewed By: qizzzh
Differential Revision: D24719910
fbshipit-source-id: d31db6304f446104ee8c7b10b902accd2919a513
