---
base_model: allenai/longformer-base-4096
tags:
- generated_from_trainer
datasets:
- essays_su_g
metrics:
- accuracy
model-index:
- name: longformer-spans
  results:
  - task:
      name: Token Classification
      type: token-classification
    dataset:
      name: essays_su_g
      type: essays_su_g
      config: spans
      split: train[80%:100%]
      args: spans
    metrics:
    - name: Accuracy
      type: accuracy
      value: 0.9313516057786306
---

<!-- This model card has been generated automatically according to the information the Trainer had access to. You
should probably proofread and complete it, then remove this comment. -->

# longformer-spans

This model is a fine-tuned version of [allenai/longformer-base-4096](https://huggingface.co/allenai/longformer-base-4096) on the essays_su_g dataset.
It achieves the following results on the evaluation set:
- Loss: 0.1886
- B: {'precision': 0.8005115089514067, 'recall': 0.900287631831256, 'f1-score': 0.8474729241877257, 'support': 1043.0}
- I: {'precision': 0.9321724709784411, 'recall': 0.9719308357348703, 'f1-score': 0.9516365688487585, 'support': 17350.0}
- O: {'precision': 0.947941598851125, 'recall': 0.8585519184912205, 'f1-score': 0.9010351495848026, 'support': 9226.0}
- Accuracy: 0.9314
- Macro avg: {'precision': 0.8935418595936575, 'recall': 0.9102567953524489, 'f1-score': 0.9000482142070956, 'support': 27619.0}
- Weighted avg: {'precision': 0.9324680497596853, 'recall': 0.9313516057786306, 'f1-score': 0.9307997762237281, 'support': 27619.0}

## 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:
- learning_rate: 2e-05
- train_batch_size: 8
- eval_batch_size: 8
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 4

### Training results

| Training Loss | Epoch | Step | Validation Loss | B                                                                                                                  | I                                                                                                                   | O                                                                                                                  | Accuracy | Macro avg                                                                                                           | Weighted avg                                                                                                        |
|:-------------:|:-----:|:----:|:---------------:|:------------------------------------------------------------------------------------------------------------------:|:-------------------------------------------------------------------------------------------------------------------:|:------------------------------------------------------------------------------------------------------------------:|:--------:|:-------------------------------------------------------------------------------------------------------------------:|:-------------------------------------------------------------------------------------------------------------------:|
| No log        | 1.0   | 41   | 0.3465          | {'precision': 0.7459016393442623, 'recall': 0.174496644295302, 'f1-score': 0.2828282828282829, 'support': 1043.0}  | {'precision': 0.8462454712392674, 'recall': 0.9827665706051874, 'f1-score': 0.90941091762447, 'support': 17350.0}   | {'precision': 0.9458898422363686, 'recall': 0.7408411012356384, 'f1-score': 0.8309020179917336, 'support': 9226.0} | 0.8714   | {'precision': 0.8460123176066329, 'recall': 0.6327014387120425, 'f1-score': 0.6743804061481621, 'support': 27619.0} | {'precision': 0.8757418451178571, 'recall': 0.8714290886708426, 'f1-score': 0.8595232027867116, 'support': 27619.0} |
| No log        | 2.0   | 82   | 0.2059          | {'precision': 0.7637130801687764, 'recall': 0.8676893576222435, 'f1-score': 0.8123877917414721, 'support': 1043.0} | {'precision': 0.9387513394619593, 'recall': 0.9593659942363112, 'f1-score': 0.9489467232975115, 'support': 17350.0} | {'precision': 0.9291049063541308, 'recall': 0.8764361586819857, 'f1-score': 0.9020023425734843, 'support': 9226.0} | 0.9282   | {'precision': 0.8771897753282888, 'recall': 0.9011638368468469, 'f1-score': 0.8877789525374893, 'support': 27619.0} | {'precision': 0.9289188728159687, 'recall': 0.9282016003475868, 'f1-score': 0.9281081765661735, 'support': 27619.0} |
| No log        | 3.0   | 123  | 0.1926          | {'precision': 0.7828618968386023, 'recall': 0.9022051773729626, 'f1-score': 0.8383073496659242, 'support': 1043.0} | {'precision': 0.9354406344242153, 'recall': 0.9654178674351584, 'f1-score': 0.950192874971636, 'support': 17350.0}  | {'precision': 0.9381976266008695, 'recall': 0.8654888358985476, 'f1-score': 0.9003777414444383, 'support': 9226.0} | 0.9296   | {'precision': 0.8855000526212291, 'recall': 0.9110372935688895, 'f1-score': 0.896292655360666, 'support': 27619.0}  | {'precision': 0.9305996331758, 'recall': 0.9296498787066875, 'f1-score': 0.9293271294770207, 'support': 27619.0}    |
| No log        | 4.0   | 164  | 0.1886          | {'precision': 0.8005115089514067, 'recall': 0.900287631831256, 'f1-score': 0.8474729241877257, 'support': 1043.0}  | {'precision': 0.9321724709784411, 'recall': 0.9719308357348703, 'f1-score': 0.9516365688487585, 'support': 17350.0} | {'precision': 0.947941598851125, 'recall': 0.8585519184912205, 'f1-score': 0.9010351495848026, 'support': 9226.0}  | 0.9314   | {'precision': 0.8935418595936575, 'recall': 0.9102567953524489, 'f1-score': 0.9000482142070956, 'support': 27619.0} | {'precision': 0.9324680497596853, 'recall': 0.9313516057786306, 'f1-score': 0.9307997762237281, 'support': 27619.0} |


### Framework versions

- Transformers 4.37.2
- Pytorch 2.2.0+cu121
- Datasets 2.17.0
- Tokenizers 0.15.2