Instructions to use riidact/ner-multilingual with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Transformers.js
How to use riidact/ner-multilingual with Transformers.js:
// npm i @huggingface/transformers import { pipeline } from '@huggingface/transformers'; // Allocate pipeline const pipe = await pipeline('token-classification', 'riidact/ner-multilingual');
| license: cc-by-nc-4.0 | |
| base_model: numind/NuNER-multilingual-v0.1 | |
| pipeline_tag: token-classification | |
| tags: | |
| - ner | |
| - multilingual | |
| - onnx | |
| - transformers.js | |
| # Riidact NER — multilingual names & places | |
| Token-classification model (PER / LOC / ORG) powering | |
| [Riidact](https://riidact.com)'s on-device name detection. Runs | |
| in the browser via transformers.js + ONNX Runtime; the quantized weights | |
| (`onnx/model_quantized.onnx`, ~178 MB) are what ships. | |
| ## License | |
| These fine-tuned weights are released under **CC-BY-NC-4.0**: free to use, | |
| share, and adapt for **non-commercial** purposes, with attribution. For commercial use, contact Riidact. | |
| The MIT-licensed base model (`numind/NuNER-multilingual-v0.1`) is unaffected | |
| and remains available under its own terms — see the NOTICE below. The | |
| non-commercial terms apply to this fine-tune, not to the underlying base. | |
| ## Coverage | |
| Fine-tuned on 103 languages; **88 are certified** by a two-tier benchmark: | |
| realistic-sentence recall with a zero-false-positive corpus (72 languages), | |
| plus held-out WikiANN validation at PER F1 ≥ 0.80 (84 languages). | |
| ## Lineage | |
| - **Backbone:** [numind/NuNER-multilingual-v0.1](https://huggingface.co/numind/NuNER-multilingual-v0.1) (MIT) | |
| - **Training data:** WikiANN (Wikipedia-derived, PAN-X annotations), 501k | |
| sentences across 103 languages, plus a small synthetic supplement of | |
| conversational name shapes (honorific patterns) that Wikipedia-derived | |
| data lacks. | |
| ## Recommended inference floors | |
| Span mean confidence, per entity: PER ≥ 0.55, LOC ≥ 0.9 (precision-first; | |
| calibrated against the benchmark corpus above). | |
| ## NOTICE | |
| This model is a derivative of `numind/NuNER-multilingual-v0.1`, licensed under | |
| the MIT License: | |
| ``` | |
| MIT License | |
| Copyright (c) NuMind | |
| Permission is hereby granted, free of charge, to any person obtaining a copy | |
| of this software and associated documentation files (the "Software"), to deal | |
| in the Software without restriction, including without limitation the rights | |
| to use, copy, modify, merge, publish, distribute, sublicense, and/or sell | |
| copies of the Software, and to permit persons to whom the Software is | |
| furnished to do so, subject to the following conditions: | |
| The above copyright notice and this permission notice shall be included in all | |
| copies or substantial portions of the Software. | |
| THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR | |
| IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, | |
| FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE | |
| AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER | |
| LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, | |
| OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE | |
| SOFTWARE. | |
| ``` | |
| The MIT terms cover the base model. The additional non-commercial terms | |
| (CC-BY-NC-4.0) apply to the fine-tuning contribution in these weights. | |