castorini / pyserini

Pyserini is a Python toolkit for reproducible information retrieval research with sparse and dense representations.
http://pyserini.io/
Apache License 2.0
1.68k stars 374 forks source link

Integrate Margin-MSE trained DistilBERT dense retrieval model #432

Closed sebastian-hofstaetter closed 3 years ago

sebastian-hofstaetter commented 3 years ago

Hi,

I just published our Margin-MSE ensemble-trained, DistilBERT-based checkpoint for dense passage retrieval here: https://huggingface.co/sebastian-hofstaetter/distilbert-dot-margin_mse-T2-msmarco

Would be cool if we could integrate that model in pyserini! I created a minimal usage example here: https://github.com/sebastian-hofstaetter/neural-ranking-kd/blob/main/minimal_bert_dot_usage_example.ipynb

Best, Sebastian

MXueguang commented 3 years ago

@lintool I think this issue can be close by https://github.com/castorini/pyserini/pull/433? CC: @sebastian-hofstaetter

lintool commented 3 years ago

👍