Support for loading jina-v5-nano retrieval adapter via eland#818
Open
gregoireco wants to merge 8 commits intomainfrom
Open
Support for loading jina-v5-nano retrieval adapter via eland#818gregoireco wants to merge 8 commits intomainfrom
gregoireco wants to merge 8 commits intomainfrom
Conversation
✅ Snyk checks have passed. No issues have been found so far.
💻 Catch issues earlier using the plugins for VS Code, JetBrains IDEs, Visual Studio, and Eclipse. |
…ntegration test - Add _JinaEmbeddingsV5WrapperModule to _TransformerTraceableModel model union type to fix mypy [arg-type] error introduced by our wrapper - Widen _SentenceTransformerWrapperModule.from_pretrained tokenizer param to accept PreTrainedTokenizerFast, fixing mypy [arg-type] at call site - Gate test_jina_v5_import_pytest.py behind TEST_JINA_INTEGRATION=1 env var so the 424 MB HuggingFace model download is skipped in CI by default Made-with: Cursor
These are optional pytorch-extras dependencies not installed in the lint environment, so mypy was failing with import-not-found errors. Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
edsavage
added a commit
to edsavage/ml-cpp
that referenced
this pull request
Mar 29, 2026
Jina Embeddings v5 is based on EuroBERT, which uses a different architecture from the BERT family: - RoPE (rotary position embeddings) → aten::sin, aten::cos - RMSNorm (instead of LayerNorm) → aten::rsqrt - SiLU activation (instead of GELU) → aten::silu Required for Eland PR elastic/eland#818 which adds support for importing Jina v5 models into Elasticsearch. Made-with: Cursor
3 tasks
edsavage
reviewed
Mar 29, 2026
Contributor
edsavage
left a comment
edsavage
left a comment
There was a problem hiding this comment.
Just a couple of comments/questions from me.
It's also woth pointing out that, due to the recent pytorch_inference hardening work, these changes now have a dependency on the ml-cpp repo. I've created a draft PR - elastic/ml-cpp#3015 - that adds the new ops to the allowlist. It will need to be merged first.
eland/ml/pytorch/transformers.py
Outdated
| if isinstance(result, (PreTrainedTokenizer, PreTrainedTokenizerFast)): | ||
| tokenizer = result | ||
| except Exception: | ||
| pass |
Contributor
There was a problem hiding this comment.
We're silently ignoring any exceptions here, should we at least log an error?
eland/ml/pytorch/nlp_ml_model.py
Outdated
| span: t.Optional[int] = None, | ||
| ): | ||
| super().__init__( | ||
| configuration_type="byte_level_bpe", |
Contributor
There was a problem hiding this comment.
Does byte_level_bpe need ES Java support?
Author
There was a problem hiding this comment.
Actually it's already supported, byte level BPE corresponds to the roberta ES tokenizer config. But I didn't understand that prior to your comment :).
edsavage
added a commit
to edsavage/ml-cpp
that referenced
this pull request
Apr 1, 2026
Jina Embeddings v5 is based on EuroBERT, which uses a different architecture from the BERT family: - RoPE (rotary position embeddings) → aten::sin, aten::cos - RMSNorm (instead of LayerNorm) → aten::rsqrt - SiLU activation (instead of GELU) → aten::silu Required for Eland PR elastic/eland#818 which adds support for importing Jina v5 models into Elasticsearch. Made-with: Cursor
2 tasks
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
4 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
No description provided.