OpenNIR issues

the effectiveness of the trained sledge-med.p model

7

Dear Sean MacAvaney, I loaded the trained model sledge-med.p following the instructions in [https://colab.research.google.com/drive/1t5UdW2Jebue1php888ldDll6yG5jQXQQ?usp=sharing] and tried to reproduce the result on the trec-covid round1 dataset. However, the output seems not...

fangguo1

Is it possible to export models?

11

Is it possible to export models so that we can use it outside of your ranking pipeline? For example BERT models fine tuned on MSMARCO.

thigm85

How to load a pretrained model?

8

nirmal2k

Getting an error about ir_measures

1

I have downloaded the ir_measures==0.2.1,when I run the following command: `scripts/pipeline.sh config/bert config/robust` an error message is occurred: ![image](https://user-images.githubusercontent.com/44088154/143192640-2c5f12ef-832c-4b96-b98b-dab4a6f1ec8f.png) Could you tell me how to fix it? Thanks

yiyaxiaozhi

getting error with vbert

4

While using vbert, I am getting the error. Please help. vbert = onir_pt.reranker('vanilla_transformer', 'bert', text_field='abstract', vocab_config={'train': True}) vbert_pipeline = ( pt.BatchRetrieve(index,wmodel='BM25',metadata=["docno", "text"]) % 1000 >>pt.text.get_text(index,"text") >>vbert ) df_res= vbert_pipeline.search("can vitamin...

somnath-banerjee

Is there any way to import customized topk training data(rather than use bm25 for first stage retrieval) ?

Can we directly train reranker model base on given retrieved data(ANCE for example)?

KkEeVvIiNnn

Problem on Loading MatchZoo Rankers

5

Perhaps it is a MatchZoo issue. Code: ```python import pyterrier as pt if not pt.started(): pt.init(tqdm='notebook') import onir_pt ``` Log: ```python --------------------------------------------------------------------------- TypeError Traceback (most recent call last) in 5...

albertoueda

MSMARCO no matching qids found

1

I tried to train conv_knrm on the MSMARCO dataset by running the following command: scripts/pipeline.sh config/conv_knrm config/msmarco However I faced "No matching QIDs found. Are you sure you are scoring...

aminbigdeli