Pavel Efimov

Hello. I've been looking for information about German Conll2003 and found your question. Official site (https://www.clips.uantwerpen.be/conll2003/ner/) mentions that organizers provide only annotation. German texts (ECI Multilingual Text Corpus) are not...

Hello @lintool. I tried reinstalling and running it several times, but nothing changed.

Hi @lintool I've found the reason of my problem. If topics comes from MS MARCO then before writing run file anserini tries to load again topic file [here](https://github.com/castorini/anserini/blob/10ae2062a5d5607657ec6abf842b08e50bf151c4/src/main/java/io/anserini/search/SearchCollection.java#L977C16-L977C138): ``` Files.newInputStream(TopicReader.getTopicPath(Path.of(Topics.MSMARCO_PASSAGE_DEV_SUBSET.path)),...

> Hope Anserini is working out for you otherwise? @lintool It's fine but not so friendly for people who have to manually download all resources: 1) Encoders: I know that...

@lintool I've just noticed that `anserini` trying to get `topics.msmarco-passage.dev-subset.txt` from the repository's root directory or trying to download it from `https://raw.githubusercontent.com/castorini/anserini-tools/master/topics-and-qrels/topics.msmarco-passage.dev-subset.txt`. But during installation I did `git submodule update...

@lintool > The onboarding path asks you do download the data separately, for pedagogical reasons I mean not the topic file which we pass as the argument. I mean the...