Hello,
I wonder if the train set top 1000 retrieval result in TREC eval format , i.e. teIn, can be released? This will allow developers to put together training sets of various sizes. Please release it in its original form, without shuffling, so truncation can be properly done.
Hello,
I wonder if the train set top 1000 retrieval result in TREC eval format , i.e. teIn, can be released? This will allow developers to put together training sets of various sizes. Please release it in its original form, without shuffling, so truncation can be properly done.