meta_data/README_s42_e4.md

metadata

base_model: allenai/longformer-base-4096
tags:
  - generated_from_trainer
datasets:
  - essays_su_g
metrics:
  - accuracy
model-index:
  - name: longformer-simple
    results:
      - task:
          name: Token Classification
          type: token-classification
        dataset:
          name: essays_su_g
          type: essays_su_g
          config: simple
          split: train[80%:100%]
          args: simple
        metrics:
          - name: Accuracy
            type: accuracy
            value: 0.8299721206415873

longformer-simple

This model is a fine-tuned version of allenai/longformer-base-4096 on the essays_su_g dataset. It achieves the following results on the evaluation set:

• Loss: 0.4321
• Claim: {'precision': 0.5835557928457021, 'recall': 0.52447216890595, 'f1-score': 0.5524387161991408, 'support': 4168.0}
• Majorclaim: {'precision': 0.6944444444444444, 'recall': 0.824814126394052, 'f1-score': 0.754035683942226, 'support': 2152.0}
• O: {'precision': 0.934596507248031, 'recall': 0.8874918707999133, 'f1-score': 0.9104353143937287, 'support': 9226.0}
• Premise: {'precision': 0.8580758203249442, 'recall': 0.8924045390540877, 'f1-score': 0.8749035689634171, 'support': 12073.0}
• Accuracy: 0.8300
• Macro avg: {'precision': 0.7676681412157805, 'recall': 0.7822956762885007, 'f1-score': 0.7729533208746281, 'support': 27619.0}
• Weighted avg: {'precision': 0.8294594932357695, 'recall': 0.8299721206415873, 'f1-score': 0.8286917107662684, '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	Claim	Majorclaim	O	Premise	Accuracy	Macro avg	Weighted avg
No log	1.0	41	0.6033	{'precision': 0.4527056753189617, 'recall': 0.2468809980806142, 'f1-score': 0.31951560316721006, 'support': 4168.0}	{'precision': 0.5835601524224279, 'recall': 0.49814126394052044, 'f1-score': 0.5374780646778641, 'support': 2152.0}	{'precision': 0.8875888965359028, 'recall': 0.8387166702796445, 'f1-score': 0.862460989745876, 'support': 9226.0}	{'precision': 0.7685754850922859, 'recall': 0.9416052348215025, 'f1-score': 0.8463371054198927, 'support': 12073.0}	0.7678	{'precision': 0.6731075523423946, 'recall': 0.6313360417805705, 'f1-score': 0.6414479407527107, 'support': 27619.0}	{'precision': 0.7462473548536118, 'recall': 0.7678409790361708, 'f1-score': 0.7481547773024915, 'support': 27619.0}
No log	2.0	82	0.4684	{'precision': 0.5774099318403116, 'recall': 0.42682341650671785, 'f1-score': 0.49082632087184436, 'support': 4168.0}	{'precision': 0.6601866251944012, 'recall': 0.7890334572490706, 'f1-score': 0.7188823031329382, 'support': 2152.0}	{'precision': 0.9429934406678593, 'recall': 0.8570344678083677, 'f1-score': 0.8979615013343932, 'support': 9226.0}	{'precision': 0.8198954421618437, 'recall': 0.9223059720036445, 'f1-score': 0.8680907460824822, 'support': 12073.0}	0.8153	{'precision': 0.7501213599661041, 'recall': 0.7487993283919501, 'f1-score': 0.7439402178554144, 'support': 27619.0}	{'precision': 0.8119780357779204, 'recall': 0.815344509214671, 'f1-score': 0.809509801603999, 'support': 27619.0}
No log	3.0	123	0.4395	{'precision': 0.5962599632127529, 'recall': 0.4666506717850288, 'f1-score': 0.5235531628532973, 'support': 4168.0}	{'precision': 0.7146464646464646, 'recall': 0.7890334572490706, 'f1-score': 0.75, 'support': 2152.0}	{'precision': 0.9242167175658862, 'recall': 0.885649252113592, 'f1-score': 0.9045220567886201, 'support': 9226.0}	{'precision': 0.8378995433789954, 'recall': 0.9119522902344074, 'f1-score': 0.873358981477809, 'support': 12073.0}	0.8264	{'precision': 0.7682556722010248, 'recall': 0.7633214178455248, 'f1-score': 0.7628585502799315, 'support': 27619.0}	{'precision': 0.820663866978074, 'recall': 0.8263876317028133, 'f1-score': 0.8213676477094007, 'support': 27619.0}
No log	4.0	164	0.4321	{'precision': 0.5835557928457021, 'recall': 0.52447216890595, 'f1-score': 0.5524387161991408, 'support': 4168.0}	{'precision': 0.6944444444444444, 'recall': 0.824814126394052, 'f1-score': 0.754035683942226, 'support': 2152.0}	{'precision': 0.934596507248031, 'recall': 0.8874918707999133, 'f1-score': 0.9104353143937287, 'support': 9226.0}	{'precision': 0.8580758203249442, 'recall': 0.8924045390540877, 'f1-score': 0.8749035689634171, 'support': 12073.0}	0.8300	{'precision': 0.7676681412157805, 'recall': 0.7822956762885007, 'f1-score': 0.7729533208746281, 'support': 27619.0}	{'precision': 0.8294594932357695, 'recall': 0.8299721206415873, 'f1-score': 0.8286917107662684, 'support': 27619.0}

Framework versions

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