software-mansion/react-native-executorch-paraphrase-multilingual-MiniLM-L12-v2

paraphrase-multilingual-MiniLM-L12-v2

This repository hosts the paraphrase-multilingual-MiniLM-L12-v2 models exported for the React Native ExecuTorch library as ExecuTorch .pte programs, ready to run on device.

Upstream model: paraphrase-multilingual-MiniLM-L12-v2

Variants

Path	Backend	Precision
coreml/paraphrase_multilingual_minilm_l12_v2_coreml_fp16.pte	coreml	fp16
vulkan/paraphrase_multilingual_minilm_l12_v2_vulkan_fp16.pte	vulkan	fp16
xnnpack/paraphrase_multilingual_minilm_l12_v2_xnnpack_fp32.pte	xnnpack	fp32
xnnpack/paraphrase_multilingual_minilm_l12_v2_xnnpack_8da4w.pte	xnnpack	8da4w

Repository structure

config.json                                                      59 B
coreml/config.json                                               971 B
coreml/paraphrase_multilingual_minilm_l12_v2_coreml_fp16.pte     225 MB
tokenizer.json                                                   16.3 MB
tokenizer_config.json                                            526 B
vulkan/config.json                                               971 B
vulkan/paraphrase_multilingual_minilm_l12_v2_vulkan_fp16.pte     224 MB
xnnpack/config.json                                              1.6 kB
xnnpack/paraphrase_multilingual_minilm_l12_v2_xnnpack_8da4w.pte  379 MB
xnnpack/paraphrase_multilingual_minilm_l12_v2_xnnpack_fp32.pte   448 MB

Compatibility

These files are published for the ExecuTorch v1.4.1 runtime. ExecuTorch gives no forward compatibility guarantee, so an older runtime may fail to load them.

To use them in React Native ExecuTorch, pass the model constant shipped in the library's model registry to the corresponding task pipeline. See the documentation.

To load these files in your own ExecuTorch runtime, read the compatibility note first.

Model details

• Architecture: 12-layer, 12-head BERT with hidden size 384 (initialized from xlm-roberta-base) + mean pooling + L2 norm. No additional dense projection head — the model output dim equals the encoder hidden size.
• Output dimension: 384.
• Max sequence length: 126 tokens (128 − 2 for the <s> / </s> wrapping; the exporter concatenates these XLM-R-style start/end tokens at id 0 / 2 inside the program).
• Vocabulary: 250 037 SentencePiece pieces.
• Languages: 50+ (multilingual).
• Typical strength: cross-lingual sentence similarity and medium-length sentence retrieval — designed for paraphrase mining and cross-lingual search. Short single-word queries in non-English languages are this model's weakest case; longer sentences and/or English inputs give markedly better ranking.

Export notes

The exporter wraps the HuggingFace transformer with the standard sentence-transformers contract: token IDs go in, the program prepends <s> and appends </s>, mean pooling is applied to the last hidden state weighted by the attention mask, and the output is L2-normalized to a 384-d vector.

Unsupported combinations (rejected by the exporter, documented for reference):

• XNNPACK + fp16 — model.to(torch.float16) causes softmax / LayerNorm overflow and the runtime output is NaN. XNNPACK's size wins come from quantization, not fp16.
• CoreML + 8da4w — coremltools has no MIL mapping for the torch.int8 tensors torchao emits (KeyError: torch.int8). The CoreML-native way to shrink further is ct.optimize.coreml palette/linear quantization, not torchao source transforms.