I have been delving into the FreshDiskANN paper and come across some points where I ask for further clarifications:
Could you confirm if the results in Figure 2 are derived from the test_concurr_merge_insert function in the diskv2 branch?
Additionally, I am interested in understanding how to use the test_concurr_merge_insert function, particularly with the SIFT1M dataset. Could you provide a brief explanation or an example for this? I am also curious about the role of the parameters in this context such as range, search_L1, search_L2.
I have been delving into the FreshDiskANN paper and come across some points where I ask for further clarifications:
Could you confirm if the results in Figure 2 are derived from the test_concurr_merge_insert function in the diskv2 branch?
Additionally, I am interested in understanding how to use the test_concurr_merge_insert function, particularly with the SIFT1M dataset. Could you provide a brief explanation or an example for this? I am also curious about the role of the parameters in this context such as range, search_L1, search_L2.