Tensor User Guide
edit page

This user guide explains the common steps needed to work with tensors in Vespa:

Setting up tensor fields in search definitions
Feeding tensors to Vespa
Querying Vespa with tensors
Ranking with tensors
Constant tensors
Tensor Java API - javadoc
Common use cases

Most of the contents discussed here is demonstrated in the tensor sample app. For a quick introduction to tensors, refer to working with tensors and the tensor reference guide. See also using TensorFlow models or using ONNX models which explains how to use TensorFlow or ONNX models directly in Vespa.

Tensor document fields

In typical use, a document contains one or more tensor fields to be used for ranking - this example sets up a tensor field called tensor_attribute:

field tensor_attribute type tensor<float>(x[4]) {
    indexing: attribute | summary
}

A tensor requires a type - x[4] means indexed dimension of size 4, where x{} means mapped dimension. tensor field in search definitions. For details on tensor types, refer to the tensor type reference.

Feeding tensors

There are two options when feeding tensors.

Convert some other field, for instance a vector of floats, to a tensor during document processing using the tensor Java API - refer to the tensor Java API.

Feed tensors using the tensor JSON format - example:

{
    "fields": {
        "tensor_attribute": {
            "cells": [
                { "address" : { "x" : "0" }, "value": 1.0 },
                { "address" : { "x" : "1" }, "value": 2.0 },
                { "address" : { "x" : "2" }, "value": 3.0 },
                { "address" : { "x" : "3" }, "value": 5.0 }
            ]
        }
    }
}

Here, the x-dimension is indexed, so the indices must be numeric. For mapped dimensions, indices can be any textual value. See the sample app.

Tensors can be updated - one can add, remove and modify tensor cells, or assign a completely new tensor value.

Querying with tensors

Tensors can only be used for ranking, not searching. The tensor can either be supplied in the query string, or constructed from some other data or data source. In the latter case, please refer to the tensor Java API for details on how to construct tensors programmatically.

Use the query(key) rank feature - here ranking.features.query(tensor) - to send a tensor with the query to ranking. Set the tensor type using a query profile type (the tensor type cannot be inferred from its value). Create a file search/query-profiles/types/root.xml in the application package - example:

<query-profile-type id="root" inherits="native">
  <field name="ranking.features.query(tensor)" type="tensor&lt;float&gt;(x[4])" />
</query-profile-type>

Also, configure the default query profile to use this this (or other query profiles as needed) - search/query-profiles/default.xml:

<query-profile id="default" type="root" />

A tensor can be constructed directly from the tensor literal form. The corresponding literal form of the tensor in the feeding section is:

{{x:0}: 1.0, {x:1}: 2.0, {x:2}: 3.0, {x:3}: 5.0}

Example query (not url-encoded for readability), using rank profile dot_product and the tensor in query(tensor):

http://host:port/search/?ranking=dot_product&ranking.features.query(tensor)={{x:0}:1.0,{x:1}:2.0,{x:2}:3.0,{x:3}:5.0}&yql=select * from sources * where sddocname contains "music";

Alternatively, use a custom searcher, and send the tensor in a parameter, like:

http://host:port/search/?tensor={{x:0}:1.0,{x:1}:2.0,{x:2}:3.0,{x:3}:5.0}&yql=select * from sources * where sddocname contains "music";

public class ExampleTensorSearcher extends Searcher {

    @Override
    public Result search(Query query, Execution execution) {
        Object tensorProperty = query.properties().get("tensor");
        if (tensorProperty != null) {
            Tensor tensor = Tensor.from("tensor<float>(x[4])", tensorProperty.toString());
            query.getRanking().getFeatures().put("query(tensor)", tensor);
            query.properties().set(new CompoundName("ranking"), "dot_product");
        }
        return execution.search(query);
    }
}

This grabs the value of the tensor query parameter, and constructs a com.yahoo.tensor.Tensor object directly from the value. It then adds this object to the query as a rank feature. You can also create the Tensor object programmatically. Refer to the tensor Java API and the tensor sample app.

Ranking with tensors

Tensors are used during ranking to modify a document's rank score given the query. Typical operations are dot products between tensors of order 1 (vectors), or matrix products between tensors of order 2 (matrices). Tensors are used in rank expressions as rank features. Two rank features are defined above:

attribute(tensor_attribute): the tensor associated with the document
query(tensor): the tensor sent with the query

These can be used in rank expressions. Note that the final rank score of a document must resolve into a single double value - example:

rank-profile dot_product {
    first-phase {
        expression: sum(query(tensor)*attribute(tensor_attribute))
    }
}

This takes the product of the query tensor and the document tensor, and sums all fields thus resolving into a single value which is used as the rank score. In the case above, the value is 39.0 (1*1 + 2*2 + 3*3 + 5*5).

There are some ranking functions that are specific for tensors:

map(tensor, f(x)(...))	Returns a new tensor with the lambda function defined in `f(x)(...)` applied to each cell.
reduce(tensor, aggregator, dim1, dim2, ...)	Returns a new tensor with the `aggregator` applied across dimensions dim1, dim2, etc. If no dimensions are specified, reduce over all dimensions.
join(tensor1, tensor2, f(x,y)(...))	Returns a new tensor constructed from the natural join between `tensor1` and `tensor2`, with the resulting cells having the value as calculated from `f(x,y)(...)`, where `x` is the cell value from `tensor1` and `y` from `tensor2`.

These primitives allow for a great deal of flexibility when combined. The above rank expression is equivalently:

rank-profile dot_product {
    first-phase {
        expression {
            reduce(
                join(
                    query(tensor),
                    attribute(tensor_attribute),
                    f(x,y)(x * y)
                ),
                sum
            )
        }
    }
}

...and represents the general dot product for tensors of any order. Details about tensor ranking functions including lambda expression and available aggregators can be found in the tensor reference documentation. More examples of tensor expression can be found in the tensor introduction.

Performance considerations

Tensor expressions are fairly concise, and since the expressions themselves are independent of the data size, the actual workload during ranking can be significant for large tensors.

When using tensors in ranking it is important to have an understanding of the potential computational cost for each query. As an example, assume the dot product of two tensors with 1000 values each, e.g. tensor<double>(x[1000]). Assuming one query tensor and one document tensor, the operation is:

sum(query(tensor1) * attribute(tensor2))

If 8 bytes is used to store each value (e.g. using a double), each tensor is approximately 8KB. With for instance a Haswell architecture the theoretical upper memory bandwidth is 68GB/s, which is around 9 million document ranking evaluations per second. With 1 million documents, this means the maximum throughput, with regards to pure memory bandwidth, is 9 queries per second (per node).

Even though you would typically not do the above without reducing the search space first (using matching and first phase), it is important to consider the memory bandwidth and other hardware limitations when developing ranking expressions with tensors.

Using a smaller value type increases performance, trading off precision: tensor<float>(x[1000]) uses 4 bytes per cell value.

Constant tensors

In addition to document tensors and query tensors, constant tensors can be put in the application package. This is useful when constant tensors are used in ranking expressions, for instance machine learned models. Example from the sample app (with a slightly changed tensor type):

constant tensor_constant {
    file: constants/constant_tensor_file.json
    type: tensor<float>(x[4])
}

This defines a new tensor rank feature with the type as defined and the contents distributed with the application package in the file constants/constant_tensor_file.json. The format of this file is the tensor JSON format, it can be compressed, see the reference for examples.

To use this tensor in a rank expression, encapsulate the constant name with constant(...):

rank-profile use_constant_tensor {
    first-phase {
        expression: sum(query(tensor) * attribute(tensor_attribute) * constant(tensor_constant))
    }
}

The above expression combines three tensors: the query tensor, the document tensor and a constant tensor.

Use cases

In the following section, find common use cases that can be solved using tensor operations.

Dot Product between query and document vectors

Assume we have a set of documents where each document contains a vector of size 4. We want to calculate the dot product between the document vectors and a vector passed down with the query and rank the results according to the dot product score.

The following sd-file defines an attribute tensor field with a tensor type that has one indexed dimension x of size 4. In addition we define a rank profile that calculates the dot product.

search example {
  document example {
    field document_vector type tensor<float>(x[4]) {
      indexing: attribute | summary
    }
  }
  rank-profile dot_product {
    first-phase {
      expression: sum(query(query_vector)*attribute(document_vector))
    }
  }
}

The tensor to pass down with query is defined in a query profile type with the same tensor type as the field in the document:

<query-profile-type id="myProfileType">
  <field name="ranking.features.query(query_vector)" type="tensor&lt;float&gt;(x[4])" />
</query-profile-type>

Example document with the vector [1.0, 2.0, 3.0, 5.0]:

[
  { "put": "id:example:example::0", "fields": {
      "document_vector" : {
        "cells": [
          { "address" : { "x" : "0" }, "value": 1.0 },
          { "address" : { "x" : "1" }, "value": 2.0 },
          { "address" : { "x" : "2" }, "value": 3.0 },
          { "address" : { "x" : "3" }, "value": 5.0 }
        ]
      }
    }
  }
]

Example query set in a searcher with the vector [1.0, 2.0, 3.0, 5.0]:

public Result search(Query query, Execution execution) {
    query.getRanking().getFeatures().put("query(query_vector)",
        Tensor.Builder.of(TensorType.fromSpec("tensor<float>(x[4])")).
        cell().label("x", 0).value(1.0).
        cell().label("x", 1).value(2.0).
        cell().label("x", 2).value(3.0).
        cell().label("x", 3).value(5.0).build());
    return execution.search(query);
}

Matrix Product between 1d vector and 2d matrix

Assume we have a 3x2 matrix represented in an attribute tensor field document_matrix with a tensor type tensor<float>(x[3],y[2]) with content:

{ {x:0,y:0}:1.0, {x:1,y:0}:3.0, {x:2,y:0}:5.0, {x:0,y:1}:7.0, {x:1,y:1}:11.0, {x:2,y:1}:13.0 }

Also assume we have 1x3 vector passed down with the query as a tensor with type tensor<float>(x[3]) with content:

{ {x:0}:1.0, {x:1}:3.0, {x:2}:5.0 }

that is set as query(query_vector) in a searcher as specified in query feature.

To calculate the matrix product between the 1x3 vector and 3x2 matrix (to get a 1x2 vector) use the following ranking expression:

sum(query(query_vector) * attribute(document_matrix),x)

This is a sparse tensor product over the shared dimension x, followed by a sum over the same dimension.

Tensor User Guide edit page