umapparams.h

Go to the documentation of this file.

/*
 * Copyright (c) 2019-2025, NVIDIA CORPORATION.
 *
 * Licensed under the Apache License, Version 2.0 (the "License");
 * you may not use this file except in compliance with the License.
 * You may obtain a copy of the License at
 *
 *     http://www.apache.org/licenses/LICENSE-2.0
 *
 * Unless required by applicable law or agreed to in writing, software
 * distributed under the License is distributed on an "AS IS" BASIS,
 * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
 * See the License for the specific language governing permissions and
 * limitations under the License.
 */
 
#pragma once
 
#include <cuml/common/callback.hpp>
#include <cuml/common/distance_type.hpp>
#include <cuml/common/logger.hpp>
 
namespace ML {
 
namespace graph_build_params {
 
struct nn_descent_params_umap {
  // not directly using cuvs::neighbors::nn_descent::index_params to distinguish UMAP-exposed NN
  // Descent parameters
  size_t graph_degree              = 64;
  size_t intermediate_graph_degree = 128;
  size_t max_iterations            = 20;
  float termination_threshold      = 0.0001;
};
 
struct graph_build_params {
  size_t overlap_factor = 2;
  size_t n_clusters = 1;
  nn_descent_params_umap nn_descent_params;
};
}  // namespace graph_build_params
 
class UMAPParams {
 public:
  enum MetricType { EUCLIDEAN, CATEGORICAL };
  enum graph_build_algo { BRUTE_FORCE_KNN, NN_DESCENT };
 
  int n_neighbors = 15;
 
  int n_components = 2;
 
  int n_epochs = 0;
 
  float learning_rate = 1.0;
 
  float min_dist = 0.1;
 
  float spread = 1.0;
 
  float set_op_mix_ratio = 1.0;
 
  float local_connectivity = 1.0;
 
  float repulsion_strength = 1.0;
 
  int negative_sample_rate = 5;
 
  float transform_queue_size = 4.0;
 
  rapids_logger::level_enum verbosity = rapids_logger::level_enum::info;
 
  float a = -1.0;
 
  float b = -1.0;
 
  float initial_alpha = 1.0;
 
  int init = 1;
 
  graph_build_algo build_algo = graph_build_algo::BRUTE_FORCE_KNN;
 
  graph_build_params::graph_build_params build_params;
 
  int target_n_neighbors = -1;
 
  MetricType target_metric = CATEGORICAL;
 
  float target_weight = 0.5;
 
  uint64_t random_state = 0;
 
  bool deterministic = true;
 
  ML::distance::DistanceType metric = ML::distance::DistanceType::L2SqrtExpanded;
 
  float p = 2.0;
 
  Internals::GraphBasedDimRedCallback* callback = nullptr;
};
 
}  // namespace ML