Include only options that control how an Onnx model is loaded in OnnxModelOptions

5 files changed, 114 insertions, 141 deletions

diff --git a/config-model/src/main/java/com/yahoo/schema/OnnxModel.java b/config-model/src/main/java/com/yahoo/schema/OnnxModel.java
index 0d2c695280e..867ffdb3960 100644
--- a/config-model/src/main/java/com/yahoo/schema/OnnxModel.java
+++ b/config-model/src/main/java/com/yahoo/schema/OnnxModel.java
@@ -143,7 +143,7 @@ public class OnnxModel extends DistributableResource implements Cloneable {
 
     public void setStatelessInterOpThreads(int interOpThreads) {
         if (interOpThreads >= 0) {
-            onnxModelOptions = onnxModelOptions.withInteropThreads(interOpThreads);
+            onnxModelOptions = onnxModelOptions.withInterOpThreads(interOpThreads);
         }
     }
 
@@ -153,7 +153,7 @@ public class OnnxModel extends DistributableResource implements Cloneable {
 
     public void setStatelessIntraOpThreads(int intraOpThreads) {
         if (intraOpThreads >= 0) {
-            onnxModelOptions = onnxModelOptions.withIntraopThreads(intraOpThreads);
+            onnxModelOptions = onnxModelOptions.withIntraOpThreads(intraOpThreads);
         }
     }
 
diff --git a/config-model/src/main/java/com/yahoo/vespa/model/container/component/BertEmbedder.java b/config-model/src/main/java/com/yahoo/vespa/model/container/component/BertEmbedder.java
index 78177fd7d57..ea3caadc23a 100644
--- a/config-model/src/main/java/com/yahoo/vespa/model/container/component/BertEmbedder.java
+++ b/config-model/src/main/java/com/yahoo/vespa/model/container/component/BertEmbedder.java
@@ -1,13 +1,12 @@
 // Copyright Vespa.ai. Licensed under the terms of the Apache 2.0 license. See LICENSE in the project root.
 package com.yahoo.vespa.model.container.component;
 
+import com.yahoo.config.ModelReference;
 import com.yahoo.config.model.deploy.DeployState;
 import com.yahoo.embedding.BertBaseEmbedderConfig;
 import com.yahoo.vespa.model.container.ApplicationContainerCluster;
 import org.w3c.dom.Element;
 
-import java.util.Optional;
-
 import static com.yahoo.embedding.BertBaseEmbedderConfig.OnnxExecutionMode;
 import static com.yahoo.embedding.BertBaseEmbedderConfig.PoolingStrategy;
 import static com.yahoo.text.XML.getChildValue;
@@ -19,40 +18,49 @@ import static com.yahoo.vespa.model.container.ContainerModelEvaluation.INTEGRATI
 public class BertEmbedder extends TypedComponent implements BertBaseEmbedderConfig.Producer {
 
     private final OnnxModelOptions onnxModelOptions;
+    private final ModelReference modelRef;
+    private final ModelReference vocabRef;
+    private final Integer maxTokens;
+    private final String transformerInputIds;
+    private final String transformerAttentionMask;
+    private final String transformerTokenTypeIds;
+    private final String transformerOutput;
+    private final Integer transformerStartSequenceToken;
+    private final Integer transformerEndSequenceToken;
+    private final String poolingStrategy;
 
     public BertEmbedder(ApplicationContainerCluster cluster, Element xml, DeployState state) {
         super("ai.vespa.embedding.BertBaseEmbedder", INTEGRATION_BUNDLE_NAME, xml);
         var model = Model.fromXml(state, xml, "transformer-model").orElseThrow();
         this.onnxModelOptions = new OnnxModelOptions(
-                Optional.of(model.modelReference()),
-                Optional.of(Model.fromXml(state, xml, "tokenizer-vocab").orElseThrow().modelReference()),
-                getChildValue(xml, "max-tokens").map(Integer::parseInt),
-                getChildValue(xml, "transformer-input-ids"),
-                getChildValue(xml, "transformer-attention-mask"),
-                getChildValue(xml, "transformer-token-type-ids"),
-                getChildValue(xml, "transformer-output"),
-                getChildValue(xml, "normalize").map(Boolean::parseBoolean),
                 getChildValue(xml, "onnx-execution-mode"),
                 getChildValue(xml, "onnx-interop-threads").map(Integer::parseInt),
                 getChildValue(xml, "onnx-intraop-threads").map(Integer::parseInt),
-                getChildValue(xml, "onnx-gpu-device").map(Integer::parseInt).map(OnnxModelOptions.GpuDevice::new),
-                getChildValue(xml, "pooling-strategy"),
-                getChildValue(xml, "transformer-start-sequence-token").map(Integer::parseInt),
-                getChildValue(xml, "transformer-end-sequence-token").map(Integer::parseInt));
+                getChildValue(xml, "onnx-gpu-device").map(Integer::parseInt).map(OnnxModelOptions.GpuDevice::new));
+        modelRef = model.modelReference();
+        vocabRef = Model.fromXml(state, xml, "tokenizer-vocab").orElseThrow().modelReference();
+        maxTokens = getChildValue(xml, "max-tokens").map(Integer::parseInt).orElse(null);
+        transformerInputIds = getChildValue(xml, "transformer-input-ids").orElse(null);
+        transformerAttentionMask = getChildValue(xml, "transformer-attention-mask").orElse(null);
+        transformerTokenTypeIds = getChildValue(xml, "transformer-token-type-ids").orElse(null);
+        transformerOutput = getChildValue(xml, "transformer-output").orElse(null);
+        transformerStartSequenceToken = getChildValue(xml, "transformer-start-sequence-token").map(Integer::parseInt).orElse(null);
+        transformerEndSequenceToken = getChildValue(xml, "transformer-end-sequence-token").map(Integer::parseInt).orElse(null);
+        poolingStrategy = getChildValue(xml, "pooling-strategy").orElse(null);
         model.registerOnnxModelCost(cluster);
     }
 
     @Override
     public void getConfig(BertBaseEmbedderConfig.Builder b) {
-        b.transformerModel(onnxModelOptions.modelRef().get()).tokenizerVocab(onnxModelOptions.vocabRef().get());
-        onnxModelOptions.maxTokens().ifPresent(b::transformerMaxTokens);
-        onnxModelOptions.transformerInputIds().ifPresent(b::transformerInputIds);
-        onnxModelOptions.transformerAttentionMask().ifPresent(b::transformerAttentionMask);
-        onnxModelOptions.transformerTokenTypeIds().ifPresent(b::transformerTokenTypeIds);
-        onnxModelOptions.transformerOutput().ifPresent(b::transformerOutput);
-        onnxModelOptions.transformerStartSequenceToken().ifPresent(b::transformerStartSequenceToken);
-        onnxModelOptions.transformerEndSequenceToken().ifPresent(b::transformerEndSequenceToken);
-        onnxModelOptions.poolingStrategy().ifPresent(value -> b.poolingStrategy(PoolingStrategy.Enum.valueOf(value)));
+        b.transformerModel(modelRef).tokenizerVocab(vocabRef);
+        if (maxTokens != null) b.transformerMaxTokens(maxTokens);
+        if (transformerInputIds != null) b.transformerInputIds(transformerInputIds);
+        if (transformerAttentionMask != null) b.transformerAttentionMask(transformerAttentionMask);
+        if (transformerTokenTypeIds != null) b.transformerTokenTypeIds(transformerTokenTypeIds);
+        if (transformerOutput != null) b.transformerOutput(transformerOutput);
+        if (transformerStartSequenceToken != null) b.transformerStartSequenceToken(transformerStartSequenceToken);
+        if (transformerEndSequenceToken != null) b.transformerEndSequenceToken(transformerEndSequenceToken);
+        if (poolingStrategy != null) b.poolingStrategy(PoolingStrategy.Enum.valueOf(poolingStrategy));
         onnxModelOptions.executionMode().ifPresent(value -> b.onnxExecutionMode(OnnxExecutionMode.Enum.valueOf(value)));
         onnxModelOptions.interOpThreads().ifPresent(b::onnxInterOpThreads);
         onnxModelOptions.intraOpThreads().ifPresent(b::onnxIntraOpThreads);
diff --git a/config-model/src/main/java/com/yahoo/vespa/model/container/component/ColBertEmbedder.java b/config-model/src/main/java/com/yahoo/vespa/model/container/component/ColBertEmbedder.java
index 33753094f69..cbae50b400c 100644
--- a/config-model/src/main/java/com/yahoo/vespa/model/container/component/ColBertEmbedder.java
+++ b/config-model/src/main/java/com/yahoo/vespa/model/container/component/ColBertEmbedder.java
@@ -7,8 +7,6 @@ import com.yahoo.embedding.ColBertEmbedderConfig;
 import com.yahoo.vespa.model.container.ApplicationContainerCluster;
 import org.w3c.dom.Element;
 
-import java.util.Optional;
-
 import static com.yahoo.embedding.ColBertEmbedderConfig.TransformerExecutionMode;
 import static com.yahoo.text.XML.getChildValue;
 import static com.yahoo.vespa.model.container.ContainerModelEvaluation.INTEGRATION_BUNDLE_NAME;
@@ -20,31 +18,43 @@ import static com.yahoo.vespa.model.container.ContainerModelEvaluation.INTEGRATI
 public class ColBertEmbedder extends TypedComponent implements ColBertEmbedderConfig.Producer {
 
     private final OnnxModelOptions onnxModelOptions;
+    private final ModelReference modelRef;
+    private final ModelReference vocabRef;
+
+    private final Integer maxQueryTokens;
+
+    private final Integer maxDocumentTokens;
+
+    private final Integer transformerStartSequenceToken;
+    private final Integer transformerEndSequenceToken;
+    private final Integer transformerMaskToken;
+    private final Integer maxTokens;
+    private final String transformerInputIds;
+    private final String transformerAttentionMask;
+
+    private final String transformerOutput;
 
     public ColBertEmbedder(ApplicationContainerCluster cluster, Element xml, DeployState state) {
         super("ai.vespa.embedding.ColBertEmbedder", INTEGRATION_BUNDLE_NAME, xml);
         var model = Model.fromXml(state, xml, "transformer-model").orElseThrow();
         this.onnxModelOptions = new OnnxModelOptions(
-                Optional.of(model.modelReference()),
-                Optional.of(Model.fromXml(state, xml, "tokenizer-model")
-                        .map(Model::modelReference)
-                        .orElseGet(() -> resolveDefaultVocab(model, state))),
-                getChildValue(xml, "max-tokens").map(Integer::parseInt),
-                getChildValue(xml, "transformer-input-ids"),
-                getChildValue(xml, "transformer-attention-mask"),
-                getChildValue(xml, "transformer-token-type-ids"),
-                getChildValue(xml, "transformer-output"),
-                getChildValue(xml, "normalize").map(Boolean::parseBoolean),
                 getChildValue(xml, "onnx-execution-mode"),
                 getChildValue(xml, "onnx-interop-threads").map(Integer::parseInt),
                 getChildValue(xml, "onnx-intraop-threads").map(Integer::parseInt),
-                getChildValue(xml, "onnx-gpu-device").map(Integer::parseInt).map(OnnxModelOptions.GpuDevice::new),
-                getChildValue(xml, "pooling-strategy"),
-                getChildValue(xml, "transformer-start-sequence-token").map(Integer::parseInt),
-                getChildValue(xml, "transformer-end-sequence-token").map(Integer::parseInt),
-                getChildValue(xml, "max-query-tokens").map(Integer::parseInt),
-                getChildValue(xml, "max-document-tokens").map(Integer::parseInt),
-                getChildValue(xml, "transformer-mask-token").map(Integer::parseInt));
+                getChildValue(xml, "onnx-gpu-device").map(Integer::parseInt).map(OnnxModelOptions.GpuDevice::new));
+        modelRef = model.modelReference();
+        vocabRef = Model.fromXml(state, xml, "tokenizer-model")
+                .map(Model::modelReference)
+                .orElseGet(() -> resolveDefaultVocab(model, state));
+        maxTokens = getChildValue(xml, "max-tokens").map(Integer::parseInt).orElse(null);
+        maxQueryTokens = getChildValue(xml, "max-query-tokens").map(Integer::parseInt).orElse(null);
+        maxDocumentTokens = getChildValue(xml, "max-document-tokens").map(Integer::parseInt).orElse(null);
+        transformerStartSequenceToken = getChildValue(xml, "transformer-start-sequence-token").map(Integer::parseInt).orElse(null);
+        transformerEndSequenceToken = getChildValue(xml, "transformer-end-sequence-token").map(Integer::parseInt).orElse(null);
+        transformerMaskToken = getChildValue(xml, "transformer-mask-token").map(Integer::parseInt).orElse(null);
+        transformerInputIds = getChildValue(xml, "transformer-input-ids").orElse(null);
+        transformerAttentionMask = getChildValue(xml, "transformer-attention-mask").orElse(null);
+        transformerOutput = getChildValue(xml, "transformer-output").orElse(null);
         model.registerOnnxModelCost(cluster);
     }
 
@@ -58,16 +68,16 @@ public class ColBertEmbedder extends TypedComponent implements ColBertEmbedderCo
 
     @Override
     public void getConfig(ColBertEmbedderConfig.Builder b) {
-        b.transformerModel(onnxModelOptions.modelRef().get()).tokenizerPath(onnxModelOptions.vocabRef().get());
-       onnxModelOptions.maxTokens().ifPresent(b::transformerMaxTokens);
-       onnxModelOptions.transformerInputIds().ifPresent(b::transformerInputIds);
-       onnxModelOptions.transformerAttentionMask().ifPresent(b::transformerAttentionMask);
-       onnxModelOptions.transformerOutput().ifPresent(b::transformerOutput);
-       onnxModelOptions.maxQueryTokens().ifPresent(b::maxQueryTokens);
-       onnxModelOptions.maxDocumentTokens().ifPresent(b::maxDocumentTokens);
-       onnxModelOptions.transformerStartSequenceToken().ifPresent(b::transformerStartSequenceToken);
-       onnxModelOptions.transformerEndSequenceToken().ifPresent(b::transformerEndSequenceToken);
-       onnxModelOptions.transformerMaskToken().ifPresent(b::transformerMaskToken);
+        b.transformerModel(modelRef).tokenizerPath(vocabRef);
+        if (maxTokens != null) b.transformerMaxTokens(maxTokens);
+        if (transformerInputIds != null) b.transformerInputIds(transformerInputIds);
+        if (transformerAttentionMask != null) b.transformerAttentionMask(transformerAttentionMask);
+        if (transformerOutput != null) b.transformerOutput(transformerOutput);
+        if (maxQueryTokens != null) b.maxQueryTokens(maxQueryTokens);
+        if (maxDocumentTokens != null) b.maxDocumentTokens(maxDocumentTokens);
+        if (transformerStartSequenceToken != null) b.transformerStartSequenceToken(transformerStartSequenceToken);
+        if (transformerEndSequenceToken != null) b.transformerEndSequenceToken(transformerEndSequenceToken);
+        if (transformerMaskToken != null) b.transformerMaskToken(transformerMaskToken);
        onnxModelOptions.executionMode().ifPresent(value -> b.transformerExecutionMode(TransformerExecutionMode.Enum.valueOf(value)));
        onnxModelOptions.interOpThreads().ifPresent(b::transformerInterOpThreads);
        onnxModelOptions.intraOpThreads().ifPresent(b::transformerIntraOpThreads);
diff --git a/config-model/src/main/java/com/yahoo/vespa/model/container/component/HuggingFaceEmbedder.java b/config-model/src/main/java/com/yahoo/vespa/model/container/component/HuggingFaceEmbedder.java
index af6febeb1e5..d1bd0dce000 100644
--- a/config-model/src/main/java/com/yahoo/vespa/model/container/component/HuggingFaceEmbedder.java
+++ b/config-model/src/main/java/com/yahoo/vespa/model/container/component/HuggingFaceEmbedder.java
@@ -7,8 +7,6 @@ import com.yahoo.embedding.huggingface.HuggingFaceEmbedderConfig;
 import com.yahoo.vespa.model.container.ApplicationContainerCluster;
 import org.w3c.dom.Element;
 
-import java.util.Optional;
-
 import static com.yahoo.embedding.huggingface.HuggingFaceEmbedderConfig.PoolingStrategy;
 import static com.yahoo.embedding.huggingface.HuggingFaceEmbedderConfig.TransformerExecutionMode;
 import static com.yahoo.text.XML.getChildValue;
@@ -21,27 +19,35 @@ import static com.yahoo.vespa.model.container.ContainerModelEvaluation.INTEGRATI
 public class HuggingFaceEmbedder extends TypedComponent implements HuggingFaceEmbedderConfig.Producer {
 
     private final OnnxModelOptions onnxModelOptions;
+    private final ModelReference modelRef;
+    private final ModelReference vocabRef;
+    private final Integer maxTokens;
+    private final String transformerInputIds;
+    private final String transformerAttentionMask;
+    private final String transformerTokenTypeIds;
+    private final String transformerOutput;
+    private final Boolean normalize;
+    private final String poolingStrategy;
 
     public HuggingFaceEmbedder(ApplicationContainerCluster cluster, Element xml, DeployState state) {
         super("ai.vespa.embedding.huggingface.HuggingFaceEmbedder", INTEGRATION_BUNDLE_NAME, xml);
         var model = Model.fromXml(state, xml, "transformer-model").orElseThrow();
-        var modelRef = model.modelReference();
         this.onnxModelOptions = new OnnxModelOptions(
-                Optional.of(modelRef),
-                Optional.of(Model.fromXml(state, xml, "tokenizer-model")
-                        .map(Model::modelReference)
-                        .orElseGet(() -> resolveDefaultVocab(model, state))),
-                getChildValue(xml, "max-tokens").map(Integer::parseInt),
-                getChildValue(xml, "transformer-input-ids"),
-                getChildValue(xml, "transformer-attention-mask"),
-                getChildValue(xml, "transformer-token-type-ids"),
-                getChildValue(xml, "transformer-output"),
-                getChildValue(xml, "normalize").map(Boolean::parseBoolean),
                 getChildValue(xml, "onnx-execution-mode"),
                 getChildValue(xml, "onnx-interop-threads").map(Integer::parseInt),
                 getChildValue(xml, "onnx-intraop-threads").map(Integer::parseInt),
-                getChildValue(xml, "onnx-gpu-device").map(Integer::parseInt).map(OnnxModelOptions.GpuDevice::new),
-                getChildValue(xml, "pooling-strategy"));
+                getChildValue(xml, "onnx-gpu-device").map(Integer::parseInt).map(OnnxModelOptions.GpuDevice::new));
+        modelRef = model.modelReference();
+        vocabRef = Model.fromXml(state, xml, "tokenizer-model")
+                .map(Model::modelReference)
+                .orElseGet(() -> resolveDefaultVocab(model, state));
+        maxTokens = getChildValue(xml, "max-tokens").map(Integer::parseInt).orElse(null);
+        transformerInputIds = getChildValue(xml, "transformer-input-ids").orElse(null);
+        transformerAttentionMask = getChildValue(xml, "transformer-attention-mask").orElse(null);
+        transformerTokenTypeIds = getChildValue(xml, "transformer-token-type-ids").orElse(null);
+        transformerOutput = getChildValue(xml, "transformer-output").orElse(null);
+        normalize = getChildValue(xml, "normalize").map(Boolean::parseBoolean).orElse(null);
+        poolingStrategy = getChildValue(xml, "pooling-strategy").orElse(null);
         model.registerOnnxModelCost(cluster);
     }
 
@@ -55,18 +61,18 @@ public class HuggingFaceEmbedder extends TypedComponent implements HuggingFaceEm
 
     @Override
     public void getConfig(HuggingFaceEmbedderConfig.Builder b) {
-        b.transformerModel(onnxModelOptions.modelRef().get()).tokenizerPath(onnxModelOptions.vocabRef().get());
-        onnxModelOptions.maxTokens().ifPresent(b::transformerMaxTokens);
-        onnxModelOptions.transformerInputIds().ifPresent(b::transformerInputIds);
-        onnxModelOptions.transformerAttentionMask().ifPresent(b::transformerAttentionMask);
-        onnxModelOptions.transformerTokenTypeIds().ifPresent(b::transformerTokenTypeIds);
-        onnxModelOptions.transformerOutput().ifPresent(b::transformerOutput);
-        onnxModelOptions.normalize().ifPresent(b::normalize);
+        b.transformerModel(modelRef).tokenizerPath(vocabRef);
+        if (maxTokens != null) b.transformerMaxTokens(maxTokens);
+        if (transformerInputIds != null) b.transformerInputIds(transformerInputIds);
+        if (transformerAttentionMask != null) b.transformerAttentionMask(transformerAttentionMask);
+        if (transformerTokenTypeIds != null) b.transformerTokenTypeIds(transformerTokenTypeIds);
+        if (transformerOutput != null) b.transformerOutput(transformerOutput);
+        if (normalize != null) b.normalize(normalize);
+        if (poolingStrategy != null) b.poolingStrategy(PoolingStrategy.Enum.valueOf(poolingStrategy));
         onnxModelOptions.executionMode().ifPresent(value -> b.transformerExecutionMode(TransformerExecutionMode.Enum.valueOf(value)));
         onnxModelOptions.interOpThreads().ifPresent(b::transformerInterOpThreads);
         onnxModelOptions.intraOpThreads().ifPresent(b::transformerIntraOpThreads);
         onnxModelOptions.gpuDevice().ifPresent(value -> b.transformerGpuDevice(value.deviceNumber()));
-        onnxModelOptions.poolingStrategy().ifPresent(value -> b.poolingStrategy(PoolingStrategy.Enum.valueOf(value)));
     }
 
 }
diff --git a/config-model/src/main/java/com/yahoo/vespa/model/container/component/OnnxModelOptions.java b/config-model/src/main/java/com/yahoo/vespa/model/container/component/OnnxModelOptions.java
index 0efb8a71fe4..6347f0dc427 100644
--- a/config-model/src/main/java/com/yahoo/vespa/model/container/component/OnnxModelOptions.java
+++ b/config-model/src/main/java/com/yahoo/vespa/model/container/component/OnnxModelOptions.java
@@ -1,86 +1,35 @@
 // Copyright Vespa.ai. Licensed under the terms of the Apache 2.0 license. See LICENSE in the project root.
 package com.yahoo.vespa.model.container.component;
 
-import com.yahoo.config.ModelReference;
-
 import java.util.Optional;
 
 /**
+ * Onnx model options that are relevant when deciding if an Onnx model needs to be reloaded. If any of the
+ * values in this class change, reload is needed.
+ *
  * @author hmusum
  */
-public record OnnxModelOptions(Optional<ModelReference> modelRef, Optional<ModelReference> vocabRef,
-                               Optional<Integer> maxTokens, Optional<String> transformerInputIds,
-                               Optional<String> transformerAttentionMask, Optional<String> transformerTokenTypeIds,
-                               Optional<String> transformerOutput, Optional<Boolean> normalize,
-                               Optional<String> executionMode, Optional<Integer> interOpThreads,
-                               Optional<Integer> intraOpThreads, Optional<GpuDevice> gpuDevice,
-                               Optional<String> poolingStrategy, Optional<Integer> transformerStartSequenceToken,
-                               Optional<Integer> transformerEndSequenceToken, Optional<Integer> maxQueryTokens,
-                               Optional<Integer> maxDocumentTokens, Optional<Integer> transformerMaskToken) {
-
-
-    public OnnxModelOptions(Optional<ModelReference> modelRef, Optional<ModelReference> vocabRef,
-                            Optional<Integer> maxTokens, Optional<String> transformerInputIds,
-                            Optional<String> transformerAttentionMask, Optional<String> transformerTokenTypeIds,
-                            Optional<String> transformerOutput, Optional<Boolean> normalize,
-                            Optional<String> executionMode, Optional<Integer> interOpThreads,
-                            Optional<Integer> intraOpThreads, Optional<GpuDevice> gpuDevice,
-                            Optional<String> poolingStrategy) {
-        this(modelRef, vocabRef, maxTokens, transformerInputIds, transformerAttentionMask, transformerTokenTypeIds,
-             transformerOutput, normalize, executionMode, interOpThreads, intraOpThreads, gpuDevice, poolingStrategy,
-             Optional.empty(), Optional.empty(), Optional.empty(), Optional.empty(), Optional.empty());
-    }
+public record OnnxModelOptions(Optional<String> executionMode, Optional<Integer> interOpThreads,
+                               Optional<Integer> intraOpThreads, Optional<GpuDevice> gpuDevice) {
 
-    public OnnxModelOptions(Optional<ModelReference> modelRef, Optional<ModelReference> vocabRef,
-                            Optional<Integer> maxTokens, Optional<String> transformerInputIds,
-                            Optional<String> transformerAttentionMask, Optional<String> transformerTokenTypeIds,
-                            Optional<String> transformerOutput, Optional<Boolean> normalize,
-                            Optional<String> executionMode, Optional<Integer> interOpThreads,
-                            Optional<Integer> intraOpThreads, Optional<GpuDevice> gpuDevice,
-                            Optional<String> poolingStrategy, Optional<Integer> transformerStartSequenceToken,
-                            Optional<Integer> transformerEndSequenceToken) {
-        this(modelRef, vocabRef, maxTokens, transformerInputIds, transformerAttentionMask, transformerTokenTypeIds,
-             transformerOutput, normalize, executionMode, interOpThreads, intraOpThreads, gpuDevice, poolingStrategy,
-             transformerStartSequenceToken, transformerEndSequenceToken, Optional.empty(), Optional.empty(), Optional.empty());
-    }
-
-    public static OnnxModelOptions empty() {
-        return new OnnxModelOptions(Optional.empty(), Optional.empty(), Optional.empty(), Optional.empty(), Optional.empty(),
-                                    Optional.empty(), Optional.empty(), Optional.empty(), Optional.empty(), Optional.empty(),
-                                    Optional.empty(), Optional.empty(), Optional.empty(), Optional.empty(), Optional.empty());
+     public static OnnxModelOptions empty() {
+        return new OnnxModelOptions(Optional.empty(), Optional.empty(), Optional.empty(), Optional.empty());
     }
 
     public OnnxModelOptions withExecutionMode(String executionMode) {
-        return new OnnxModelOptions(modelRef, vocabRef, maxTokens, transformerInputIds, transformerAttentionMask,
-                                    transformerTokenTypeIds, transformerOutput, normalize, Optional.ofNullable(executionMode),
-                                    interOpThreads, intraOpThreads, gpuDevice, poolingStrategy,
-                                    transformerStartSequenceToken, transformerEndSequenceToken,
-                                    maxQueryTokens, maxDocumentTokens, transformerMaskToken);
+        return new OnnxModelOptions(Optional.ofNullable(executionMode), interOpThreads, intraOpThreads, gpuDevice);
     }
 
-    public OnnxModelOptions withInteropThreads(Integer interopThreads) {
-        return new OnnxModelOptions(modelRef, vocabRef, maxTokens, transformerInputIds, transformerAttentionMask,
-                                    transformerTokenTypeIds, transformerOutput, normalize, executionMode,
-                                    Optional.ofNullable(interopThreads), intraOpThreads, gpuDevice, poolingStrategy,
-                                    transformerStartSequenceToken, transformerEndSequenceToken,
-                                    maxQueryTokens, maxDocumentTokens, transformerMaskToken);
+    public OnnxModelOptions withInterOpThreads(Integer interOpThreads) {
+        return new OnnxModelOptions(executionMode, Optional.ofNullable(interOpThreads), intraOpThreads, gpuDevice);
     }
 
-    public OnnxModelOptions withIntraopThreads(Integer intraopThreads) {
-        return new OnnxModelOptions(modelRef, vocabRef, maxTokens, transformerInputIds, transformerAttentionMask,
-                                    transformerTokenTypeIds, transformerOutput, normalize, executionMode,
-                                    interOpThreads, Optional.ofNullable(intraopThreads), gpuDevice, poolingStrategy,
-                                    transformerStartSequenceToken, transformerEndSequenceToken,
-                                    maxQueryTokens, maxDocumentTokens, transformerMaskToken);
+    public OnnxModelOptions withIntraOpThreads(Integer intraOpThreads) {
+        return new OnnxModelOptions(executionMode, interOpThreads, Optional.ofNullable(intraOpThreads), gpuDevice);
     }
 
-
     public OnnxModelOptions withGpuDevice(GpuDevice gpuDevice) {
-        return new OnnxModelOptions(modelRef, vocabRef, maxTokens, transformerInputIds, transformerAttentionMask,
-                                    transformerTokenTypeIds, transformerOutput, normalize, executionMode,
-                                    interOpThreads, intraOpThreads, Optional.ofNullable(gpuDevice), poolingStrategy,
-                                    transformerStartSequenceToken, transformerEndSequenceToken,
-                                    maxQueryTokens, maxDocumentTokens, transformerMaskToken);
+        return new OnnxModelOptions(executionMode, interOpThreads, intraOpThreads, Optional.ofNullable(gpuDevice));
     }
 
     public record GpuDevice(int deviceNumber, boolean required) {