Li Yongkang
Li Yongkang
Hello, have you solved your problem?I want to know how to deal with data sets downloaded from the Internet and how to split them up.
@mumuyanyan
> Thx for your interest. I have fixed the issue in the mat file for a while. I think you are probably using an old version of the mat file....
> @luyug I think I have figure this prolem out, thanks. but during my experiment , I found that the loss is very difficult to converge, here is my log:...
Hi all, I am facing the same issue: the speed is very slow. I also observed that my GPU memory was only a few GB in use, even though I...
I’ve created a [code repository](https://github.com/liyongkang123/extended_beir_datasets) to deal with this. Anyone who needs it can use it.
Hi Nandan, I’m not sure if this is the right approach. If there are cases where a `doc_id `is the same as a `query_id `in the corpus, I wonder whether...
Hi Nandan, I also found the following difference: ``` retriever = EvaluateRetrieval(model, score_function=score_function, k_values=[1, 5, 10, 50, 100, 1000]) # "dot" for dot product cos_sim results = retriever.encode_and_retrieve(corpus, queries, encode_output_path=embedding_save_path,...
Hi Nandan, I can try doing it this way. However, I'm not sure when the normalization is performed. We have two strategies: 1. Perform normalization before saving the embedding, then...
Thanks, I submitted a pull request #161 here, and I hope it can be merged asap.