Blindsight third-party licences
blindsight.io/platform/legal/third-party
This document covers the Blindsight platform (the console, the desktop apps and the APIs). The blindsight.io website has its own privacy policy.
1. What this page lists
The models, tokenizer and datasets behind the on-device names model, with their creators, licences, sources and the changes Blindsight made. The model files are downloaded by the desktop app after installation. The weights and the tokenizer file are licensed under CC BY-SA 4.0; the base model and the datasets keep the licences named below. The libraries and programs inside the installers are listed in the licence notices supplied with the software.
2. Models, tokenizer and datasets
| Name | Creator | Licence | Source | What Blindsight changed |
|---|---|---|---|---|
| Weights of the on-device names model (versions 4 and 4-fp16) | Blindsight Technologies AG | CC BY-SA 4.0 | Downloaded by the desktop app after installation; not inside the installers. | Fine-tuned from the base model; converted to ONNX; INT8 (dynamic, per channel) on Apple silicon, FP16 on all other devices. The tokenizer file is the tokenizer of the base model. |
| distilbert-base-multilingual-cased (base model) | Sanh et al. | Apache-2.0 | https://huggingface.co/distilbert/distilbert-base-multilingual-cased | Weights fine-tuned by Blindsight for token classification (names, organisations, places and dates), then converted to ONNX with reduced numeric precision (FP16 or INT8). |
| Universal NER v1 (treebanks English-EWT, Portuguese-Bosque, Chinese-GSD, Chinese-GSDSimp) | Mayhew et al. | CC BY-SA 4.0 | https://www.universalner.org | Converted to a token-classification format and relabelled to person, organisation, place and date; entity classes mapped; some sentences case-changed or had names swapped. |
| Europeana Newspapers NER corpora | Europeana Newspapers Project | CC0 | https://github.com/EuropeanaNewspapers/ner-corpora | Converted to a token-classification format and relabelled to person, organisation, place and date; entity classes mapped; some sentences case-changed or had names swapped. |
| ANERcorp, CAMeL Lab train/test splits | Benajiba, Rosso and Benedi Ruiz; Obeid et al. | CC BY-SA 4.0 | https://camel.abudhabi.nyu.edu/anercorp/ | Converted to a token-classification format and relabelled to person, organisation, place and date; entity classes mapped; some sentences case-changed or had names swapped. |
| AQMAR Arabic Wikipedia Named Entity Corpus | Mohit, Schneider, Bhowmick, Oflazer and Smith | CC BY-SA 3.0 | https://www.cs.cmu.edu/~ark/ArabicNER/ | Converted to a token-classification format and relabelled to person, organisation, place and date; entity classes mapped; some sentences case-changed or had names swapped. CC BY-SA 3.0 section 4(b) permits an adaptation to be distributed under a later version of the licence with the same licence elements, here CC BY-SA 4.0. |
| GermEval 2014 NER shared task data | Benikova, Biemann, Reznicek | CC BY 4.0 | https://sites.google.com/site/germeval2014ner/ | Converted to a token-classification format and relabelled to person, organisation, place and date; entity classes mapped; some sentences case-changed or had names swapped. |
| WikiNER and wikigold | Nothman, Ringland, Radford, Murphy, Curran; Balasuriya et al. | CC BY 4.0 | https://doi.org/10.6084/m9.figshare.5462500 | Converted to a token-classification format and relabelled to person, organisation, place and date; entity classes mapped; some sentences case-changed or had names swapped. |
| MultiCoNER v2 | Fetahu, Chen, Kar, Rokhlenko, Malmasi | CC BY 4.0 | https://multiconer.github.io | Converted to a token-classification format and relabelled to person, organisation, place and date; entity classes mapped; some sentences case-changed or had names swapped. |
| LUMII AiLab FullStack, Latvian named entities | LUMII AiLab | CC BY-SA 4.0 | https://github.com/LUMII-AILab/FullStack | Converted to a token-classification format and relabelled to person, organisation, place and date; entity classes mapped; some sentences case-changed or had names swapped. |
| Wikidata name labels | Wikidata | CC0 | https://www.wikidata.org | Used to fill synthetic sentences. |
| Organisation name list (orgs-v1.json) | Wikidata | CC0 | https://www.wikidata.org | Labels and identifiers only, no Wikipedia text. A separate file, not covered by the CC BY-SA 4.0 grant for the model files. |
| Davlan/distilbert-base-multilingual-cased-ner-hrl (earlier name model, versions 1 and 2-fp16 only) | Davlan (the model card names no copyright holder) | AFL-3.0 | https://huggingface.co/Davlan/distilbert-base-multilingual-cased-ner-hrl | Converted to ONNX; the Apple silicon file is quantised to int8 and the file for every other device is converted to fp16. Weights otherwise unmodified. |
Blindsight also generated sentences by script from its own templates and lists, and had a language model write synthetic sentences with the names filled in by script. These are Blindsight's own material. No data from any Blindsight customer was used.
3. Full notice shipped with the model files
This is the notice that comes with the INT8 model files, followed by the one change that differs in the FP16 files. It includes the Apache License 2.0 text for the base model.
The Blindsight on-device names model
This model detects person names, organisations, places and dates in text on the
device. It was trained by Blindsight from the base model below on the datasets
below. The model weights and the tokenizer file in this directory are licensed
under the Creative Commons Attribution-ShareAlike 4.0 International licence
(CC BY-SA 4.0): https://creativecommons.org/licenses/by-sa/4.0/legalcode
Base model
distilbert-base-multilingual-cased, Apache License 2.0
https://huggingface.co/distilbert/distilbert-base-multilingual-cased
Sanh et al., DistilBERT, 2019. The Apache-2.0 notice of the base model is
preserved here, and the full text of the Apache License 2.0 is reproduced at
the end of this file. The base model was modified: its weights were
fine-tuned by Blindsight for token classification (names, organisations,
places and dates) and the result was converted to ONNX, with reduced numeric
precision (FP16 or INT8) in some variants.
Training data and attribution
Universal NER v1 (CC BY-SA 4.0). Mayhew et al., Universal NER: A Gold-Standard
Multilingual Named Entity Recognition Benchmark, NAACL 2024.
https://www.universalner.org
Treebanks used: English-EWT, Portuguese-Bosque, Chinese-GSD and
Chinese-GSDSimp (Universal Dependencies treebanks, CC BY-SA 4.0, with the
Universal NER annotations).
Europeana Newspapers NER corpora (CC0). Europeana Newspapers Project,
2012-2015. https://github.com/EuropeanaNewspapers/ner-corpora
ANERcorp, CAMeL Lab train/test splits (CC BY-SA 4.0). Benajiba, Rosso and
Benedi Ruiz, 2007; Obeid et al., CAMeL Tools, LREC 2020.
https://camel.abudhabi.nyu.edu/anercorp/
AQMAR Arabic Wikipedia Named Entity Corpus (CC BY-SA 3.0). Mohit, Schneider,
Bhowmick, Oflazer and Smith, Recall-Oriented Learning of Named Entities in
Arabic Wikipedia, EACL 2012. https://www.cs.cmu.edu/~ark/ArabicNER/
CC BY-SA 3.0 section 4(b) permits an adaptation to be distributed under a
later version of the licence with the same licence elements, here CC BY-SA
4.0.
GermEval 2014 NER shared task data (CC BY 4.0). Benikova, Biemann, Reznicek,
2014. https://sites.google.com/site/germeval2014ner/
WikiNER and wikigold (CC BY 4.0). Nothman, Ringland, Radford, Murphy, Curran,
Learning multilingual named entity recognition from Wikipedia, Artificial
Intelligence 2013; Balasuriya et al., 2009.
https://doi.org/10.6084/m9.figshare.5462500
MultiCoNER v2 (CC BY 4.0). Fetahu, Chen, Kar, Rokhlenko, Malmasi, 2023.
https://multiconer.github.io
LUMII AiLab FullStack, Latvian named entities (CC BY-SA 4.0).
https://github.com/LUMII-AILab/FullStack
Wikidata name labels (CC0), used to fill synthetic sentences.
https://www.wikidata.org
Several of the corpora above (AQMAR, GermEval 2014, WikiNER, wikigold and
MultiCoNER v2) consist of sentences from Wikipedia, which is licensed under
CC BY-SA.
Round 6 adds sentences generated by Blindsight scripts from its own
templates, and same-type name swaps inside the corpora above, filled with
the Wikidata labels above. Round 8 adds sentences generated by Blindsight
scripts from its own templates (JSON, YAML, tables, logs, e-mail headers,
citations, chat exports, tool calls and short questions that contain names),
filled with Wikidata name labels and corpus name frequencies from the
corpora above. Round 9 adds sentences generated by Blindsight scripts from its
own templates (multi-record logs, CSV cells, user= lines, mail headers,
headings, usernames and bot names kept as non-names, organisation classes,
first names with surnames that are also common nouns), and same-type
organisation swaps inside the real training sentences of the corpora above
(German, English, Arabic, French). No new third-party data.
Synthetic sentences written with a language model, with names,
organisations, places and dates filled in by script, so every label is
exact.
Further sentences (government bodies, product and tool names, questions)
generated by script from lists written by Blindsight, and real training
sentences of the corpora above with organisations swapped by script.
Round 11 adds, by script from lists written by Blindsight, organisation
names of companies, newspapers and magazines, venues and clubs, universities
and schools, parties and international bodies in German and Portuguese
sentences (templates, and real training sentences of the corpora above with
an organisation swapped), further German, Austrian and Swiss business and
public-body names, persons with fresh first names and surnames in many
frames, and further non-name examples. No new third-party data.
Round 12 adds, by script from lists written by Blindsight and the Wikidata
name labels above: person names in English, German, Spanish, Italian, Dutch,
French and Portuguese (first names after labels and in signatures, surnames,
initials, public figures, placeholder names, titled and fictional names, and
common words kept as non-names); organisation names (public bodies and
agencies, armed forces and social-security bodies, department and team nouns,
sports leagues, clubs and venues, newspapers and journals, universities,
schools, libraries and museums, small companies with legal forms, all-capital
company names, company names in code, configuration and e-mail text); further
German, Austrian and Swiss organisation classes; and further non-name
examples. No new third-party data.
Round 13 adds, by script from lists written by Blindsight and the Wikidata
name labels above: person names in English, German, Spanish, Italian, Dutch,
French and Portuguese (lone first names after verbs of address and in
sign-offs, surnames without a title, initial forms, public figures and
stage names in ordinary questions, placeholder names in templates, titled
names, and capitalised common words, role words, assistant and tool names
kept as non-names), and replays, unchanged, person-bearing training
sentences of the corpora above (including ANERcorp, AQMAR, GermEval 2014,
Europeana, WikiNER, MultiCoNER v2 and Universal NER) that were also used in
earlier rounds. No new third-party data. Round 13 also adds, by script from
lists written by Blindsight, non-name examples (occupation words that are
also surnames, plural colour nouns, postcodes with city names). No new
third-party data.
These datasets were converted to a token-classification format and relabelled
to person, organisation, place and date; entity classes were mapped; some
sentences were case-changed or had names swapped.
Provided as is without warranties, as set out in the disclaimer of warranties
of the licences named above.
Based on distilbert-base-multilingual-cased (Apache-2.0), fine-tuned by
Blindsight Technologies AG.
Modification statement (this variant: dynamic per-channel INT8)
model.onnx in this directory was converted to ONNX and INT8 quantised by
Blindsight Technologies AG from the weights, which Blindsight
Technologies AG fine-tuned from distilbert-base-multilingual-cased
(https://huggingface.co/distilbert/distilbert-base-multilingual-cased).
tokenizer.json is the tokenizer of that base model.
Organisation list
Organisation names are taken from Wikidata (CC0); names belong to their
owners. The file orgs-v1.json holds labels and identifiers only, no
Wikipedia text (see NOTICE-orgs-v1.txt). It is a separate file, not
covered by the CC BY-SA 4.0 grant above, and the "license" field of
manifest.json does not apply to it.
Integrity checks are not access restrictions
The model files are shipped as separate, unencrypted files. Code signing
and the sha256 values in manifest.json only let the Blindsight app check
that the files are intact; they do not restrict copying, modifying or
sharing these files under their licences.
No data from any Blindsight customer was used.
Apache License 2.0 (applies to the base model distilbert-base-multilingual-cased)
------------------------------------------------------------------------
Apache License
Version 2.0, January 2004
http://www.apache.org/licenses/
TERMS AND CONDITIONS FOR USE, REPRODUCTION, AND DISTRIBUTION
1. Definitions.
"License" shall mean the terms and conditions for use, reproduction,
and distribution as defined by Sections 1 through 9 of this document.
"Licensor" shall mean the copyright owner or entity authorized by
the copyright owner that is granting the License.
"Legal Entity" shall mean the union of the acting entity and all
other entities that control, are controlled by, or are under common
control with that entity. For the purposes of this definition,
"control" means (i) the power, direct or indirect, to cause the
direction or management of such entity, whether by contract or
otherwise, or (ii) ownership of fifty percent (50%) or more of the
outstanding shares, or (iii) beneficial ownership of such entity.
"You" (or "Your") shall mean an individual or Legal Entity
exercising permissions granted by this License.
"Source" form shall mean the preferred form for making modifications,
including but not limited to software source code, documentation
source, and configuration files.
"Object" form shall mean any form resulting from mechanical
transformation or translation of a Source form, including but
not limited to compiled object code, generated documentation,
and conversions to other media types.
"Work" shall mean the work of authorship, whether in Source or
Object form, made available under the License, as indicated by a
copyright notice that is included in or attached to the work
(an example is provided in the Appendix below).
"Derivative Works" shall mean any work, whether in Source or Object
form, that is based on (or derived from) the Work and for which the
editorial revisions, annotations, elaborations, or other modifications
represent, as a whole, an original work of authorship. For the purposes
of this License, Derivative Works shall not include works that remain
separable from, or merely link (or bind by name) to the interfaces of,
the Work and Derivative Works thereof.
"Contribution" shall mean any work of authorship, including
the original version of the Work and any modifications or additions
to that Work or Derivative Works thereof, that is intentionally
submitted to Licensor for inclusion in the Work by the copyright owner
or by an individual or Legal Entity authorized to submit on behalf of
the copyright owner. For the purposes of this definition, "submitted"
means any form of electronic, verbal, or written communication sent
to the Licensor or its representatives, including but not limited to
communication on electronic mailing lists, source code control systems,
and issue tracking systems that are managed by, or on behalf of, the
Licensor for the purpose of discussing and improving the Work, but
excluding communication that is conspicuously marked or otherwise
designated in writing by the copyright owner as "Not a Contribution."
"Contributor" shall mean Licensor and any individual or Legal Entity
on behalf of whom a Contribution has been received by Licensor and
subsequently incorporated within the Work.
2. Grant of Copyright License. Subject to the terms and conditions of
this License, each Contributor hereby grants to You a perpetual,
worldwide, non-exclusive, no-charge, royalty-free, irrevocable
copyright license to reproduce, prepare Derivative Works of,
publicly display, publicly perform, sublicense, and distribute the
Work and such Derivative Works in Source or Object form.
3. Grant of Patent License. Subject to the terms and conditions of
this License, each Contributor hereby grants to You a perpetual,
worldwide, non-exclusive, no-charge, royalty-free, irrevocable
(except as stated in this section) patent license to make, have made,
use, offer to sell, sell, import, and otherwise transfer the Work,
where such license applies only to those patent claims licensable
by such Contributor that are necessarily infringed by their
Contribution(s) alone or by combination of their Contribution(s)
with the Work to which such Contribution(s) was submitted. If You
institute patent litigation against any entity (including a
cross-claim or counterclaim in a lawsuit) alleging that the Work
or a Contribution incorporated within the Work constitutes direct
or contributory patent infringement, then any patent licenses
granted to You under this License for that Work shall terminate
as of the date such litigation is filed.
4. Redistribution. You may reproduce and distribute copies of the
Work or Derivative Works thereof in any medium, with or without
modifications, and in Source or Object form, provided that You
meet the following conditions:
(a) You must give any other recipients of the Work or
Derivative Works a copy of this License; and
(b) You must cause any modified files to carry prominent notices
stating that You changed the files; and
(c) You must retain, in the Source form of any Derivative Works
that You distribute, all copyright, patent, trademark, and
attribution notices from the Source form of the Work,
excluding those notices that do not pertain to any part of
the Derivative Works; and
(d) If the Work includes a "NOTICE" text file as part of its
distribution, then any Derivative Works that You distribute must
include a readable copy of the attribution notices contained
within such NOTICE file, excluding those notices that do not
pertain to any part of the Derivative Works, in at least one
of the following places: within a NOTICE text file distributed
as part of the Derivative Works; within the Source form or
documentation, if provided along with the Derivative Works; or,
within a display generated by the Derivative Works, if and
wherever such third-party notices normally appear. The contents
of the NOTICE file are for informational purposes only and
do not modify the License. You may add Your own attribution
notices within Derivative Works that You distribute, alongside
or as an addendum to the NOTICE text from the Work, provided
that such additional attribution notices cannot be construed
as modifying the License.
You may add Your own copyright statement to Your modifications and
may provide additional or different license terms and conditions
for use, reproduction, or distribution of Your modifications, or
for any such Derivative Works as a whole, provided Your use,
reproduction, and distribution of the Work otherwise complies with
the conditions stated in this License.
5. Submission of Contributions. Unless You explicitly state otherwise,
any Contribution intentionally submitted for inclusion in the Work
by You to the Licensor shall be under the terms and conditions of
this License, without any additional terms or conditions.
Notwithstanding the above, nothing herein shall supersede or modify
the terms of any separate license agreement you may have executed
with Licensor regarding such Contributions.
6. Trademarks. This License does not grant permission to use the trade
names, trademarks, service marks, or product names of the Licensor,
except as required for reasonable and customary use in describing the
origin of the Work and reproducing the content of the NOTICE file.
7. Disclaimer of Warranty. Unless required by applicable law or
agreed to in writing, Licensor provides the Work (and each
Contributor provides its Contributions) on an "AS IS" BASIS,
WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or
implied, including, without limitation, any warranties or conditions
of TITLE, NON-INFRINGEMENT, MERCHANTABILITY, or FITNESS FOR A
PARTICULAR PURPOSE. You are solely responsible for determining the
appropriateness of using or redistributing the Work and assume any
risks associated with Your exercise of permissions under this License.
8. Limitation of Liability. In no event and under no legal theory,
whether in tort (including negligence), contract, or otherwise,
unless required by applicable law (such as deliberate and grossly
negligent acts) or agreed to in writing, shall any Contributor be
liable to You for damages, including any direct, indirect, special,
incidental, or consequential damages of any character arising as a
result of this License or out of the use or inability to use the
Work (including but not limited to damages for loss of goodwill,
work stoppage, computer failure or malfunction, or any and all
other commercial damages or losses), even if such Contributor
has been advised of the possibility of such damages.
9. Accepting Warranty or Additional Liability. While redistributing
the Work or Derivative Works thereof, You may choose to offer,
and charge a fee for, acceptance of support, warranty, indemnity,
or other liability obligations and/or rights consistent with this
License. However, in accepting such obligations, You may act only
on Your own behalf and on Your sole responsibility, not on behalf
of any other Contributor, and only if You agree to indemnify,
defend, and hold each Contributor harmless for any liability
incurred by, or claims asserted against, such Contributor by reason
of your accepting any such warranty or additional liability.
END OF TERMS AND CONDITIONS
APPENDIX: How to apply the Apache License to your work.
To apply the Apache License to your work, attach the following
boilerplate notice, with the fields enclosed by brackets "[]"
replaced with your own identifying information. (Don't include
the brackets!) The text should be enclosed in the appropriate
comment syntax for the file format. We also recommend that a
file or class name and description of purpose be included on the
same "printed page" as the copyright notice for easier
identification within third-party archives.
Copyright [yyyy] [name of copyright owner]
Licensed under the Apache License, Version 2.0 (the "License");
you may not use this file except in compliance with the License.
You may obtain a copy of the License at
http://www.apache.org/licenses/LICENSE-2.0
Unless required by applicable law or agreed to in writing, software
distributed under the License is distributed on an "AS IS" BASIS,
WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
See the License for the specific language governing permissions and
limitations under the License.
Modification statement in the FP16 notice
Modification statement (this variant: FP16) model.onnx in this directory was converted to ONNX and FP16 quantised by Blindsight Technologies AG from the weights, which Blindsight Technologies AG fine-tuned from distilbert-base-multilingual-cased (https://huggingface.co/distilbert/distilbert-base-multilingual-cased). tokenizer.json is the tokenizer of that base model.
Organisation list notice
Organisation name list orgs-v1: names and identifiers from Wikidata (https://www.wikidata.org), released under CC0 1.0 Universal (public domain dedication). No attribution is required; given for transparency.