You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
{{ message }}
Repository navigation
[License Question] UIE Nano model weights / vocab / config redistribution and ONNX conversion
#11252
I am developing an open-source desktop application called SafePrompt, and currently use the uie-nano model from PaddleNLP 2.6.1 Taskflow for local PERSON / ORG information extraction.
Before publicly distributing the Windows binary, I would like to clarify the license and redistribution terms for the UIE Nano model assets.
I understand that the PaddleNLP source repository is licensed under Apache-2.0, but I have not found an official statement clearly specifying the license for these model weights, vocabulary, and configuration assets.
Could you please clarify:
What license applies to the uie-nano model weights?
Do vocab.txt, tokenizer_config.json, and config.json use the same license?
Is redistribution of these assets together with a third-party desktop application allowed?
Is conversion of the Paddle model weights to ONNX format permitted?
Is redistribution of the converted ONNX model permitted?
If permitted, may the converted ONNX model, vocabulary, and config files be bundled in a publicly distributed Windows Portable / ZIP application?
Are there any additional requirements, such as:
including the Apache-2.0 LICENSE
NOTICE / attribution
model source attribution
modification / format-conversion notice
any other redistribution obligations
The ONNX model in our project is only a format conversion of the official UIE Nano model. We have not retrained the model or intentionally changed its model semantics.
This clarification is required for our public release license gate, especially regarding:
redistribution of model weights
model format conversion
redistribution of converted model
redistribution of vocab/config assets
If these assets are covered by Apache-2.0, it would be very helpful if you could explicitly confirm that Apache-2.0 covers the model weights and related assets listed above.
reacted with thumbs up emoji reacted with thumbs down emoji reacted with laugh emoji reacted with hooray emoji reacted with confused emoji reacted with heart emoji reacted with rocket emoji reacted with eyes emoji
Uh oh!
There was an error while loading. Please reload this page.
Hello PaddleNLP team,
Thank you for releasing PaddleNLP and UIE.
I am developing an open-source desktop application called SafePrompt, and currently use the
uie-nanomodel from PaddleNLP 2.6.1 Taskflow for local PERSON / ORG information extraction.Before publicly distributing the Windows binary, I would like to clarify the license and redistribution terms for the UIE Nano model assets.
The official assets currently used are:
UIE Nano weights
https://bj.bcebos.com/paddlenlp/taskflow/information_extraction/uie_nano_v1.1/model_state.pdparams
UIE Nano config
https://bj.bcebos.com/paddlenlp/taskflow/information_extraction/uie_nano/config.json
UIE Base vocab
https://bj.bcebos.com/paddlenlp/taskflow/information_extraction/uie_base/vocab.txt
UIE Base tokenizer config
https://bj.bcebos.com/paddlenlp/taskflow/information_extraction/uie_base/tokenizer_config.json
I understand that the PaddleNLP source repository is licensed under Apache-2.0, but I have not found an official statement clearly specifying the license for these model weights, vocabulary, and configuration assets.
Could you please clarify:
What license applies to the
uie-nanomodel weights?Do
vocab.txt,tokenizer_config.json, andconfig.jsonuse the same license?Is redistribution of these assets together with a third-party desktop application allowed?
Is conversion of the Paddle model weights to ONNX format permitted?
Is redistribution of the converted ONNX model permitted?
If permitted, may the converted ONNX model, vocabulary, and config files be bundled in a publicly distributed Windows Portable / ZIP application?
Are there any additional requirements, such as:
The ONNX model in our project is only a format conversion of the official UIE Nano model. We have not retrained the model or intentionally changed its model semantics.
This clarification is required for our public release license gate, especially regarding:
If these assets are covered by Apache-2.0, it would be very helpful if you could explicitly confirm that Apache-2.0 covers the model weights and related assets listed above.
Thank you very much.
All reactions