Closed slz929 closed 2 years ago
Yes, precomputed_matches are the retrieval results.
I used the same command that you have copied here python main.py -em 0 -sm 0 -dn corridor -ddir datasets/corridor/ -mdir precomputed_matches/corridor/ -techs NetVLAD_Precomputed for the clone of this remote repository but it works fine for me. I cannot reproduce this error on my end.
Yes, precomputed_matches are the retrieval results.
I used the same command that you have copied here _python main.py -em 0 -sm 0 -dn corridor -ddir datasets/corridor/ -mdir precomputed_matches/corridor/ -techs NetVLADPrecomputed for the clone of this remote repository but it works fine for me. I cannot reproduce this error on my end.
fine,maybe blame on my python environments, Thanks !
I want to use the matching data we had already computed for netVLAD or CALC:
(And I wonder what you mean about the 'matching data' ? Is that the precomputed retrieval results? )
python main.py -em 0 -sm 0 -dn corridor -ddir datasets/corridor/ -mdir precomputed_matches/corridor/ -techs NetVLAD_Precomputed
but errors like below:
File "/VPR-Bench/performance_comparison.py", line 256, in compute_matches if (retr in ground_truth_info[itr][1]): IndexError: index 111 is out of bounds for axis 0 with size 111