diff options
| author | Bjørn Christian Seime <bjorncs@yahooinc.com> | 2023-05-30 16:52:45 +0200 |
|---|---|---|
| committer | Bjørn Christian Seime <bjorncs@yahooinc.com> | 2023-05-30 16:52:47 +0200 |
| commit | bafc5758799fa2e9e2ece36a85ea27c2c92982fd (patch) | |
| tree | ba55b18d5a1d9b7e48209b18f659a47db6d77f23 /model-integration | |
| parent | 4279a9423629c5ea9daa15d5d297f2347b7c22ec (diff) | |
Properly ignore token type ids from tokenizer if disabled
Diffstat (limited to 'model-integration')
| -rw-r--r-- | model-integration/src/main/java/ai/vespa/embedding/huggingface/HuggingFaceEmbedder.java | 4 |
1 files changed, 2 insertions, 2 deletions
diff --git a/model-integration/src/main/java/ai/vespa/embedding/huggingface/HuggingFaceEmbedder.java b/model-integration/src/main/java/ai/vespa/embedding/huggingface/HuggingFaceEmbedder.java index 88028da85c1..01804656bb6 100644 --- a/model-integration/src/main/java/ai/vespa/embedding/huggingface/HuggingFaceEmbedder.java +++ b/model-integration/src/main/java/ai/vespa/embedding/huggingface/HuggingFaceEmbedder.java @@ -84,11 +84,11 @@ public class HuggingFaceEmbedder extends AbstractComponent implements Embedder { var encoding = tokenizer.encode(s, context.getLanguage()); Tensor inputSequence = createTensorRepresentation(encoding.ids(), "d1"); Tensor attentionMask = createTensorRepresentation(encoding.attentionMask(), "d1"); - Tensor tokenTypeIds = createTensorRepresentation(encoding.typeIds(), "d1"); + Tensor tokenTypeIds = tokenTypeIdsName.isEmpty() ? null : createTensorRepresentation(encoding.typeIds(), "d1"); Map<String, Tensor> inputs; - if (tokenTypeIds.isEmpty()) { + if (tokenTypeIdsName.isEmpty() || tokenTypeIds.isEmpty()) { inputs = Map.of(inputIdsName, inputSequence.expand("d0"), attentionMaskName, attentionMask.expand("d0")); } else { |