Inside Milvus 1.1.0

Milvus v1.1.0 adds the capability to specify partition names in `get_entity_by_id()` and `delete_entity_by_id()` calls, optimizing operations by restricting searches to relevant partitions. This significantly improves performance for large datasets. Additionally, a new `release_collection()` method allows manual unloading of collections from memory to free up resources. Performance improvements were also noted for `get_entity_by_id()`, Hnswlib index building, and IVF index training.