Transformers
PyTorch
Safetensors
English
bart
text2text-generation
seq2seq
relation-extraction
triple-generation
entity-linking
entity-type-linking
relation-linking
Eval Results (legacy)
Instructions to use ibm-research/knowgl-large with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Transformers
How to use ibm-research/knowgl-large with Transformers:
# pip install -U transformers accelerate # Load model directly from transformers import AutoTokenizer, AutoModelForSeq2SeqLM tokenizer = AutoTokenizer.from_pretrained("ibm-research/knowgl-large") model = AutoModelForSeq2SeqLM.from_pretrained("ibm-research/knowgl-large", device_map="auto") - Notebooks
- Google Colab
- Kaggle
Commit ·
dbef85d
1
Parent(s): d1b9887
Update README.md
Browse files
README.md
CHANGED
|
@@ -28,11 +28,11 @@ license: cc-by-nc-sa-4.0
|
|
| 28 |
|
| 29 |
# KnowGL: Knowledge Generation and Linking from Text
|
| 30 |
|
| 31 |
-
The `knowgl-large` model
|
| 32 |
```
|
| 33 |
[(subject mention # subject label # subject type) | relation label | (object mention # object label # object type)]
|
| 34 |
```
|
| 35 |
-
If there is more than one triple generated, they are separated by `$` in the output.
|
| 36 |
|
| 37 |
The model achieves state-of-the-art results for relation extraction on the REBEL dataset. See results in [Mihindukulasooriya et al. (ISWC 2022)](https://arxiv.org/pdf/2207.05188.pdf).
|
| 38 |
|
|
|
|
| 28 |
|
| 29 |
# KnowGL: Knowledge Generation and Linking from Text
|
| 30 |
|
| 31 |
+
The `knowgl-large` model is trained by combining Wikidata with an extended version of the training data [REBEL](https://huggingface.co/datasets/Babelscape/rebel-dataset) dataset. Given a sentence, KnowGL generates triple(s) in the following format:
|
| 32 |
```
|
| 33 |
[(subject mention # subject label # subject type) | relation label | (object mention # object label # object type)]
|
| 34 |
```
|
| 35 |
+
If there is more than one triple generated, they are separated by `$` in the output. More details in [Rossiello et al. (AAAI 2023)](https://arxiv.org/pdf/2210.13952.pdf).
|
| 36 |
|
| 37 |
The model achieves state-of-the-art results for relation extraction on the REBEL dataset. See results in [Mihindukulasooriya et al. (ISWC 2022)](https://arxiv.org/pdf/2207.05188.pdf).
|
| 38 |
|