Jimmy Lin
Jimmy Lin
Actually, I'm thinking that each index would get its own separate underlying searcher instance with its own thread pool. So, all of these would be completely independent... ``` http://localhost:8081/api/v1.0/indexes/msmarco-v2.1-doc-segmented-shard00.arctic-embed-l.hnsw-int8/ http://localhost:8081/api/v1.0/indexes/msmarco-v2.1-doc-segmented-shard01.arctic-embed-l.hnsw-int8/...
I will write up in a guide shortly, but just to pass along. How to run TREC RAG24 test queries with ArcticEmbed-L shards: ```bash SHARDS=(00 01 02 03 04 05...
What about `ShardInfo` to parallel `IndexInfo`? ``` MSMARCO_V21_DOC_SEGMENTED_ARCTIC_EMBED_L_HNSW_INT8( "msmarco-v2.1-doc-segmented.arctic-embed-l.hnsw-int8", .... new IndexInfo[] { MSMARCO_V21_DOC_SEGMENTED_SHARD00_ARCTIC_EMBED_L_HNSW_INT8, MSMARCO_V21_DOC_SEGMENTED_SHARD01_ARCTIC_EMBED_L_HNSW_INT8 ... } ... ) ```
Remember to use `-optimize` to merge down to a single index segment.
Closed by #1764 and #1770
If topic keys are integers, then it would be nice to have int sort order, like 1, 2, 3 ... 10, 11... otherwise, we get string sort order, like 1,...
Yes, `isdigit()` makes sense. @crystina-z - related this one: https://github.com/castorini/pyserini/pull/1289 Actually @ChuanMeng - would you be willing to do a PR to fix? I would like to expand the contributor...
#1349 was merged. closing.
> I also agree that (1) is cleaner, but I'm not sure if it works for BEIR (and correct me if I'm wrong) since I don't think .flat is a...
Adding `/api/v1/...` is a breaking change, but I like the idea. Let's do `v1.0` instead of just `v1`. I'm okay with all the other aspects of your proposal. Let's do...