metadata
license: cc-by-nc-sa-4.0
library_name: transformers
pipeline_tag: token-classification
widget:
- text: 'Do you think that looks like a cat? Answer: I don''t think so.'
- example_title: cat
xlm-roberta-base for token classification, specifically fine-tuned for question-answer extraction for English
This is the xlm-roberta-base, fine-tuned on manually annotated Finnish data and ChatGPT-annotated data.
Hyperparameters
batch_size = 8
epochs = 10 (trained for less)
base_LM_model = "xlm-roberta-base"
max_seq_len = 512
learning_rate = 5e-5
Performance
Accuracy = 0.88
Question F1 = 0.77
Answer F1 = 0.81
Usage
To get the best question-answer pairs use the huggingface pipeline with no aggregation strategy and do some post-processing like in this script.