Ruth's picture
Upload README.md
f87b83e
|
raw
history blame
1.6 kB
metadata
language:
  - de
license: mit
datasets:
  - germeval_14
metrics:
  - precision
  - recall
  - f1
  - accuracy
model-index:
  - name: gelectra-large-germeval_14
    results:
      - task:
          name: Token Classification
          type: token-classification
        dataset:
          name: germeval_14
          type: germeval_14
          args: default
        metrics:
          - name: precision
            type: precision
            value: 0.85778125349123
          - name: recall
            type: recall
            value: 0.8761839552664613
          - name: f1
            type: f1
            value: 0.8668849497572543
          - name: accuracy
            type: accuracy
            value: 0.976306430788049

gelectra-large-germeval_14

This model is a fine-tuned version of deepset/gelectra-large on the germeval_14 dataset. It achieves the following results on the evaluation set:

  • precision: 0.8578
  • recall: 0.8762
  • f1: 0.8669
  • accuracy: 0.9763

Model description

More information needed

Intended uses & limitations

More information needed

Training and evaluation data

More information needed

Training procedure

Training hyperparameters

The following hyperparameters were used during training:

  • num_train_epochs: 5
  • train_batch_size: 8
  • eval_batch_size: 8
  • learning_rate: 2e-05
  • weight_decay_rate: 0.01
  • num_warmup_steps: 0
  • fp16: True

Framework versions

  • Transformers 4.18.0
  • Datasets 1.18.0
  • Tokenizers 0.12.1