Instructions to use tner/roberta-large-ontonotes5 with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Transformers
How to use tner/roberta-large-ontonotes5 with Transformers:
# Use a pipeline as a high-level helper from transformers import pipeline pipe = pipeline("token-classification", model="tner/roberta-large-ontonotes5")# Load model directly from transformers import AutoTokenizer, AutoModelForTokenClassification tokenizer = AutoTokenizer.from_pretrained("tner/roberta-large-ontonotes5") model = AutoModelForTokenClassification.from_pretrained("tner/roberta-large-ontonotes5", device_map="auto") - Notebooks
- Google Colab
- Kaggle
License clarification for commercial use and redistribution
#3
by hfiguera - opened
Hello,
I am developing a library for PII detection and anonymization and am evaluatingtner/roberta-large-ontonotes5 as an optional local NER model.
The checkpoint revision evaluated is:
0bce50f7884d5bb040469c907c897d4b061ccbb4
The library will not bundle or redistribute the model. Before deciding whether
this model can be recommended for production use, I need to understand its
licensing accurately.
The model card does not currently declare a license. I found that:
- the T-NER software repository is MIT licensed;
- the
roberta-largebase model is marked MIT; - this checkpoint was fine-tuned using
tner/ontonotes5; - OntoNotes 5.0 is distributed under an LDC agreement.
Could you please clarify:
- What license covers the model checkpoint weights?
- What license covers the tokenizer and configuration files?
- Is commercial inference with this checkpoint permitted?
- May downstream users redistribute or mirror the checkpoint?
- Do OntoNotes/LDC restrictions apply to users who only use the checkpoint
and never access or redistribute the OntoNotes dataset? - Are attribution or notice requirements applicable?
- Did the training organization have permission to publish the checkpoint for
these uses?
If the checkpoint is intended to be MIT licensed, would you be willing to add
an SPDX license: mit field and a model-specific license or notice file to
this repository?
Thank you.