Skip to content

feat: add PARAPHRASE_MULTILINGUAL_MINILM_L12_V2 text embeddings model #1115

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
msluszniak merged 3 commits into from
May 7, 2026
Merged

feat: add PARAPHRASE_MULTILINGUAL_MINILM_L12_V2 text embeddings model #1115

Show file tree
Hide file tree
Changes from 2 commits
Commits
Show all changes
3 commits
Select commit Hold shift + click to select a range
689ce63
feat: add PARAPHRASE_MULTILINGUAL_MINILM_L12_V2 text embeddings model
msluszniak Apr 30, 2026
2cfefc9
chore: trim to single XNNPACK 8da4w variant + add benchmark rows
msluszniak May 6, 2026
1d9ea9a
refactor: rename paraphrase-multilingual export to _QUANTIZED suffix
msluszniak May 7, 2026
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
5 changes: 5 additions & 0 deletions apps/text-embeddings/app/text-embeddings/index.tsx
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -20,6 +20,7 @@ import {
MULTI_QA_MPNET_BASE_DOT_V1,
DISTILUSE_BASE_MULTILINGUAL_CASED_V2_8DA4W,
DISTILUSE_BASE_MULTILINGUAL_CASED_V2_COREML,
PARAPHRASE_MULTILINGUAL_MINILM_L12_V2_8DA4W,
TextEmbeddingsProps,
} from 'react-native-executorch';

Expand All @@ -38,6 +39,10 @@ const MODELS: { label: string; value: TextEmbeddingModel }[] = [
label: 'Multilingual DistilUSE (CoreML)',
value: DISTILUSE_BASE_MULTILINGUAL_CASED_V2_COREML,
},
{
label: 'Multilingual Paraphrase (8da4w)',
value: PARAPHRASE_MULTILINGUAL_MINILM_L12_V2_8DA4W,
},
];
import { useIsFocused } from '@react-navigation/native';
import { dotProduct } from '../../utils/math';
Expand Down
19 changes: 10 additions & 9 deletions docs/docs/02-benchmarks/inference-time.md
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -180,15 +180,16 @@ Average time to synthesize speech from an input text of approximately 60 tokens,
Benchmark times for text embeddings are highly dependent on the sentence length. The numbers below are based on a sentence of around 80 tokens. For shorter or longer sentences, inference time may vary accordingly.
:::

| Model / Device | iPhone 17 Pro [ms] | OnePlus 12 [ms] |
| ---------------------------------------------------- | :----------------: | :-------------: |
| ALL_MINILM_L6_V2 (XNNPACK) | 7 | 21 |
| ALL_MPNET_BASE_V2 (XNNPACK) | 24 | 90 |
| MULTI_QA_MINILM_L6_COS_V1 (XNNPACK) | 7 | 19 |
| MULTI_QA_MPNET_BASE_DOT_V1 (XNNPACK) | 24 | 88 |
| CLIP_VIT_BASE_PATCH32_TEXT (XNNPACK) | 14 | 39 |
| DISTILUSE_BASE_MULTILINGUAL_CASED_V2 (XNNPACK 8da4w) | 16 | 15 |
| DISTILUSE_BASE_MULTILINGUAL_CASED_V2 (Core ML FP32) | 15 | - |
| Model / Device | iPhone 17 Pro [ms] | OnePlus 12 [ms] |
| ----------------------------------------------------- | :----------------: | :-------------: |
| ALL_MINILM_L6_V2 (XNNPACK) | 7 | 21 |
| ALL_MPNET_BASE_V2 (XNNPACK) | 24 | 90 |
| MULTI_QA_MINILM_L6_COS_V1 (XNNPACK) | 7 | 19 |
| MULTI_QA_MPNET_BASE_DOT_V1 (XNNPACK) | 24 | 88 |
| CLIP_VIT_BASE_PATCH32_TEXT (XNNPACK) | 14 | 39 |
| DISTILUSE_BASE_MULTILINGUAL_CASED_V2 (XNNPACK 8da4w) | 16 | 15 |
| DISTILUSE_BASE_MULTILINGUAL_CASED_V2 (Core ML FP32) | 15 | - |
| PARAPHRASE_MULTILINGUAL_MINILM_L12_V2 (XNNPACK 8da4w) | 14 | 15 |

## Image Embeddings

Expand Down
19 changes: 10 additions & 9 deletions docs/docs/02-benchmarks/memory-usage.md
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -98,15 +98,16 @@ The reported memory usage values include the memory footprint of the Phonemis pa

## Text Embeddings

| Model / Device | iPhone 17 Pro [MB] | OnePlus 12 [MB] |
| ---------------------------------------------------- | :----------------: | :-------------: |
| ALL_MINILM_L6_V2 (XNNPACK) | 110 | 95 |
| ALL_MPNET_BASE_V2 (XNNPACK) | 455 | 405 |
| MULTI_QA_MINILM_L6_COS_V1 (XNNPACK) | 140 | 120 |
| MULTI_QA_MPNET_BASE_DOT_V1 (XNNPACK) | 455 | 435 |
| CLIP_VIT_BASE_PATCH32_TEXT (XNNPACK) | 280 | 200 |
| DISTILUSE_BASE_MULTILINGUAL_CASED_V2 (XNNPACK 8da4w) | 36 | 44 |
| DISTILUSE_BASE_MULTILINGUAL_CASED_V2 (Core ML FP32) | 55 | - |
| Model / Device | iPhone 17 Pro [MB] | OnePlus 12 [MB] |
| ----------------------------------------------------- | :----------------: | :-------------: |
| ALL_MINILM_L6_V2 (XNNPACK) | 110 | 95 |
| ALL_MPNET_BASE_V2 (XNNPACK) | 455 | 405 |
| MULTI_QA_MINILM_L6_COS_V1 (XNNPACK) | 140 | 120 |
| MULTI_QA_MPNET_BASE_DOT_V1 (XNNPACK) | 455 | 435 |
| CLIP_VIT_BASE_PATCH32_TEXT (XNNPACK) | 280 | 200 |
| DISTILUSE_BASE_MULTILINGUAL_CASED_V2 (XNNPACK 8da4w) | 36 | 44 |
| DISTILUSE_BASE_MULTILINGUAL_CASED_V2 (Core ML FP32) | 55 | - |
| PARAPHRASE_MULTILINGUAL_MINILM_L12_V2 (XNNPACK 8da4w) | 131 | 141 |

## Image Embeddings

Expand Down
1 change: 1 addition & 0 deletions docs/docs/02-benchmarks/model-size.md
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -128,6 +128,7 @@ title: Model Size
| CLIP_VIT_BASE_PATCH32_TEXT | 254 |
| DISTILUSE_BASE_MULTILINGUAL_CASED_V2_8DA4W | 393 |
| DISTILUSE_BASE_MULTILINGUAL_CASED_V2_COREML | 541 |
| PARAPHRASE_MULTILINGUAL_MINILM_L12_V2_8DA4W | 397 |

## Image Embeddings

Expand Down
17 changes: 9 additions & 8 deletions docs/docs/03-hooks/01-natural-language-processing/useTextEmbeddings.md
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -101,14 +101,15 @@ function App() {

## Supported models

| Model | Language | Max Tokens | Embedding Dimensions | Description |
| ------------------------------------------------------------------------------------------------------------------------- | :-----------: | :--------: | :------------------: | -------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
| [all-MiniLM-L6-v2](https://huggingface.co/sentence-transformers/all-MiniLM-L6-v2) | English | 254 | 384 | All-round model tuned for many use-cases. Trained on a large and diverse dataset of over 1 billion training pairs. |
| [all-mpnet-base-v2](https://huggingface.co/sentence-transformers/all-mpnet-base-v2) | English | 382 | 768 | All-round model tuned for many use-cases. Trained on a large and diverse dataset of over 1 billion training pairs. |
| [multi-qa-MiniLM-L6-cos-v1](https://huggingface.co/sentence-transformers/multi-qa-MiniLM-L6-cos-v1) | English | 509 | 384 | This model was tuned for semantic search: Given a query/question, it can find relevant passages. It was trained on a large and diverse set of (question, answer) pairs. |
| [multi-qa-mpnet-base-dot-v1](https://huggingface.co/sentence-transformers/multi-qa-mpnet-base-dot-v1) | English | 510 | 768 | This model was tuned for semantic search: Given a query/question, it can find relevant passages. It was trained on a large and diverse set of (question, answer) pairs. |
| [distiluse-base-multilingual-cased-v2](https://huggingface.co/sentence-transformers/distiluse-base-multilingual-cased-v2) | 50+ languages | 126 | 512 | Multilingual DistilBERT with a 768→512 projection head. Recommended when broader language coverage matters more than the exact English quality of MiniLM/MPNet. |
| [clip-vit-base-patch32-text](https://huggingface.co/openai/clip-vit-base-patch32) | English | 74 | 512 | CLIP (Contrastive Language-Image Pre-Training) is a neural network trained on a variety of (image, text) pairs. CLIP allows to embed images and text into the same vector space. This allows to find similar images as well as to implement image search. This is the text encoder part of the CLIP model. To embed images checkout [clip-vit-base-patch32-image](../02-computer-vision/useImageEmbeddings.md#supported-models). |
| Model | Language | Max Tokens | Embedding Dimensions | Description |
| --------------------------------------------------------------------------------------------------------------------------- | :-----------: | :--------: | :------------------: | -------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
| [all-MiniLM-L6-v2](https://huggingface.co/sentence-transformers/all-MiniLM-L6-v2) | English | 254 | 384 | All-round model tuned for many use-cases. Trained on a large and diverse dataset of over 1 billion training pairs. |
| [all-mpnet-base-v2](https://huggingface.co/sentence-transformers/all-mpnet-base-v2) | English | 382 | 768 | All-round model tuned for many use-cases. Trained on a large and diverse dataset of over 1 billion training pairs. |
| [multi-qa-MiniLM-L6-cos-v1](https://huggingface.co/sentence-transformers/multi-qa-MiniLM-L6-cos-v1) | English | 509 | 384 | This model was tuned for semantic search: Given a query/question, it can find relevant passages. It was trained on a large and diverse set of (question, answer) pairs. |
| [multi-qa-mpnet-base-dot-v1](https://huggingface.co/sentence-transformers/multi-qa-mpnet-base-dot-v1) | English | 510 | 768 | This model was tuned for semantic search: Given a query/question, it can find relevant passages. It was trained on a large and diverse set of (question, answer) pairs. |
| [distiluse-base-multilingual-cased-v2](https://huggingface.co/sentence-transformers/distiluse-base-multilingual-cased-v2) | 50+ languages | 126 | 512 | Multilingual DistilBERT with a 768→512 projection head. Recommended when broader language coverage matters more than the exact English quality of MiniLM/MPNet. |
| [paraphrase-multilingual-MiniLM-L12-v2](https://huggingface.co/sentence-transformers/paraphrase-multilingual-MiniLM-L12-v2) | 50+ languages | 126 | 384 | Multilingual MiniLM-L12 distilled from paraphrase-multilingual-mpnet-base-v2. Compact (≈118 M params) sentence encoder for cross-lingual semantic similarity and retrieval across 50+ languages. |
| [clip-vit-base-patch32-text](https://huggingface.co/openai/clip-vit-base-patch32) | English | 74 | 512 | CLIP (Contrastive Language-Image Pre-Training) is a neural network trained on a variety of (image, text) pairs. CLIP allows to embed images and text into the same vector space. This allows to find similar images as well as to implement image search. This is the text encoder part of the CLIP model. To embed images checkout [clip-vit-base-patch32-image](../02-computer-vision/useImageEmbeddings.md#supported-models). |

**`Max Tokens`** - The maximum number of tokens that can be processed by the model. If the input text exceeds this limit, it will be truncated.

Expand Down
12 changes: 12 additions & 0 deletions packages/react-native-executorch/src/constants/modelUrls.ts
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -1102,6 +1102,8 @@ const MULTI_QA_MPNET_BASE_DOT_V1_TOKENIZER = `${URL_PREFIX}-multi-qa-mpnet-base-
const DISTILUSE_BASE_MULTILINGUAL_CASED_V2_8DA4W_MODEL = `${URL_PREFIX}-distiluse-base-multilingual-cased-v2/${NEXT_VERSION_TAG}/xnnpack/distiluse-base-multilingual-cased-v2_xnnpack_8da4w.pte`;
const DISTILUSE_BASE_MULTILINGUAL_CASED_V2_COREML_MODEL = `${URL_PREFIX}-distiluse-base-multilingual-cased-v2/${NEXT_VERSION_TAG}/coreml/distiluse-base-multilingual-cased-v2_coreml_fp32.pte`;
const DISTILUSE_BASE_MULTILINGUAL_CASED_V2_TOKENIZER = `${URL_PREFIX}-distiluse-base-multilingual-cased-v2/${NEXT_VERSION_TAG}/tokenizer.json`;
const PARAPHRASE_MULTILINGUAL_MINILM_L12_V2_8DA4W_MODEL = `${URL_PREFIX}-paraphrase-multilingual-MiniLM-L12-v2/${NEXT_VERSION_TAG}/xnnpack/paraphrase-multilingual-MiniLM-L12-v2_xnnpack_8da4w.pte`;
const PARAPHRASE_MULTILINGUAL_MINILM_L12_V2_TOKENIZER = `${URL_PREFIX}-paraphrase-multilingual-MiniLM-L12-v2/${NEXT_VERSION_TAG}/tokenizer.json`;
const CLIP_VIT_BASE_PATCH32_TEXT_MODEL = `${URL_PREFIX}-clip-vit-base-patch32/${VERSION_TAG}/xnnpack/clip_vit_base_patch32_text_xnnpack_fp32.pte`;
const CLIP_VIT_BASE_PATCH32_TEXT_TOKENIZER = `${URL_PREFIX}-clip-vit-base-patch32/${VERSION_TAG}/tokenizer.json`;

Expand Down Expand Up @@ -1159,6 +1161,15 @@ export const DISTILUSE_BASE_MULTILINGUAL_CASED_V2_COREML = {
tokenizerSource: DISTILUSE_BASE_MULTILINGUAL_CASED_V2_TOKENIZER,
} as const;

/**
* @category Models - Text Embeddings
*/
export const PARAPHRASE_MULTILINGUAL_MINILM_L12_V2_8DA4W = {

@barhanc barhanc May 6, 2026

Copy link
Copy Markdown
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I think for now it should be PARAPHRASE_MULTILINGUAL_MINILM_L12_V2_QUANTIZED so that we maintain consistent naming (and we should probably retroactively change the DISTILUSE_BASE_MULTILINGUAL_CASED_V2_8DA4W also). We had some discussion about model registry and naming (#612) so we can change the symbols all in one go when implementing that.

@chmjkb chmjkb May 7, 2026

Copy link
Copy Markdown
Collaborator

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

agreed, I think that answers your question @msluszniak

@msluszniak msluszniak May 7, 2026

Copy link
Copy Markdown
Member Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Ok, so I'll change the name for this model and for distiluse we will fix this in one holistic change.

modelName: 'paraphrase-multilingual-minilm-l12-v2-8da4w',
modelSource: PARAPHRASE_MULTILINGUAL_MINILM_L12_V2_8DA4W_MODEL,
tokenizerSource: PARAPHRASE_MULTILINGUAL_MINILM_L12_V2_TOKENIZER,
} as const;

/**
* @category Models - Text Embeddings
*/
Expand Down Expand Up @@ -1349,6 +1360,7 @@ export const MODEL_REGISTRY = {
MULTI_QA_MPNET_BASE_DOT_V1,
DISTILUSE_BASE_MULTILINGUAL_CASED_V2_8DA4W,
DISTILUSE_BASE_MULTILINGUAL_CASED_V2_COREML,
PARAPHRASE_MULTILINGUAL_MINILM_L12_V2_8DA4W,
CLIP_VIT_BASE_PATCH32_TEXT,
BK_SDM_TINY_VPRED_512,
BK_SDM_TINY_VPRED_256,
Expand Down
1 change: 1 addition & 0 deletions packages/react-native-executorch/src/types/textEmbeddings.ts
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -12,6 +12,7 @@ export type TextEmbeddingsModelName =
| 'multi-qa-mpnet-base-dot-v1'
| 'distiluse-base-multilingual-cased-v2-8da4w'
| 'distiluse-base-multilingual-cased-v2-coreml'
| 'paraphrase-multilingual-minilm-l12-v2-8da4w'
| 'clip-vit-base-patch32-text';

/**
Expand Down
Loading