diff --git a/python/raft-ann-bench/src/raft-ann-bench/run/conf/datasets.yml b/python/raft-ann-bench/src/raft-ann-bench/run/conf/datasets.yml new file mode 100644 index 0000000000..bffa632815 --- /dev/null +++ b/python/raft-ann-bench/src/raft-ann-bench/run/conf/datasets.yml @@ -0,0 +1,85 @@ +- "name": "bigann-100M", + "base_file": "bigann-1B/base.1B.u8bin", + "subset_size": 100000000, + "query_file": "bigann-1B/query.public.10K.u8bin", + "groundtruth_neighbors_file": "bigann-100M/groundtruth.neighbors.ibin", + "distance": "euclidean" + +- "name": "deep-1B", + "base_file": "deep-1B/base.1B.fbin", + "query_file": "deep-1B/query.public.10K.fbin", + "groundtruth_neighbors_file": "deep-1B/groundtruth.neighbors.ibin", + "distance": "inner_product" + +- "name": "bigann-100M", + "base_file": "bigann-1B/base.1B.u8bin", + "subset_size": 100000000, + "query_file": "bigann-1B/query.public.10K.u8bin", + "groundtruth_neighbors_file": "bigann-100M/groundtruth.neighbors.ibin", + "distance": "euclidean" + +- "name": "deep-image-96-inner", + "base_file": "deep-image-96-inner/base.fbin", + "query_file": "deep-image-96-inner/query.fbin", + "groundtruth_neighbors_file": "deep-image-96-inner/groundtruth.neighbors.ibin", + "distance": "euclidean" + +- "name": "fashion-mnist-784-euclidean", + "base_file": "fashion-mnist-784-euclidean/base.fbin", + "query_file": "fashion-mnist-784-euclidean/query.fbin", + "groundtruth_neighbors_file": "fashion-mnist-784-euclidean/groundtruth.neighbors.ibin", + "distance": "euclidean" + +- "name": "gist-960-euclidean", + "base_file": "gist-960-euclidean/base.fbin", + "query_file": "gist-960-euclidean/query.fbin", + "distance": "euclidean" + +- "name": "glove-50-angular", + "base_file": "glove-50-angular/base.fbin", + "query_file": "glove-50-angular/query.fbin", + "distance": "euclidean" + +- "name": "glove-50-inner", + "base_file": "glove-50-inner/base.fbin", + "query_file": "glove-50-inner/query.fbin", + "distance": "euclidean" + +- "name": "glove-100-angular", + "base_file": "glove-100-angular/base.fbin", + "query_file": "glove-100-angular/query.fbin", + "distance": "euclidean" + +- "name": "glove-100-inner", + "base_file": "glove-100-inner/base.fbin", + "query_file": "glove-100-inner/query.fbin", + "distance": "euclidean" + +- "name": "lastfm-65-angular", + "base_file": "lastfm-65-angular/base.fbin", + "query_file": "lastfm-65-angular/query.fbin", + "distance": "euclidean" + +- "name": "mnist-784-euclidean", + "base_file": "mnist-784-euclidean/base.fbin", + "query_file": "mnist-784-euclidean/query.fbin", + "groundtruth_neighbors_file": "mnist-784-euclidean/groundtruth.neighbors.ibin", + "distance": "euclidean" + +- "name": "nytimes-256-angular", + "base_file": "nytimes-256-angular/base.fbin", + "query_file": "nytimes-256-angular/query.fbin", + "groundtruth_neighbors_file": "nytimes-256-angular/groundtruth.neighbors.ibin", + "distance": "euclidean" + +- "name": "nytimes-256-inner", + "base_file": "nytimes-256-inner/base.fbin", + "query_file": "nytimes-256-inner/query.fbin", + "groundtruth_neighbors_file": "nytimes-256-inner/groundtruth.neighbors.ibin", + "distance": "euclidean" + +- "name": "sift-128-euclidean", + "base_file": "sift-128-euclidean/base.fbin", + "query_file": "sift-128-euclidean/query.fbin", + "groundtruth_neighbors_file": "sift-128-euclidean/groundtruth.neighbors.ibin", + "distance": "euclidean" \ No newline at end of file