longformer-full_labels

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

• Loss: 0.3818
• B-claim: {'precision': 0.5588235294117647, 'recall': 0.46830985915492956, 'f1-score': 0.5095785440613027, 'support': 284.0}
• B-majorclaim: {'precision': 0.8787878787878788, 'recall': 0.20567375886524822, 'f1-score': 0.3333333333333333, 'support': 141.0}
• B-premise: {'precision': 0.7287735849056604, 'recall': 0.8728813559322034, 'f1-score': 0.794344473007712, 'support': 708.0}
• I-claim: {'precision': 0.6021926389976507, 'recall': 0.5673880964092474, 'f1-score': 0.5842725085475498, 'support': 4066.0}
• I-majorclaim: {'precision': 0.7885196374622356, 'recall': 0.7767857142857143, 'f1-score': 0.782608695652174, 'support': 2016.0}
• I-premise: {'precision': 0.8760707709550877, 'recall': 0.8973349733497334, 'f1-score': 0.8865753868589484, 'support': 12195.0}
• O: {'precision': 0.9648159446817165, 'recall': 0.9631509491422191, 'f1-score': 0.9639827279654559, 'support': 9851.0}
• Accuracy: 0.8573
• Macro avg: {'precision': 0.7711405693145706, 'recall': 0.6787892438770422, 'f1-score': 0.693527952775211, 'support': 29261.0}
• Weighted avg: {'precision': 0.8552285449410628, 'recall': 0.8572502648576603, 'f1-score': 0.8549088561404111, 'support': 29261.0}

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: 5

Training results

Training Loss	Epoch	Step	Validation Loss	B-claim	B-majorclaim	B-premise	I-claim	I-majorclaim	I-premise	O	Accuracy	Macro avg	Weighted avg
No log	1.0	41	0.7363	{'precision': 0.0, 'recall': 0.0, 'f1-score': 0.0, 'support': 284.0}	{'precision': 0.0, 'recall': 0.0, 'f1-score': 0.0, 'support': 141.0}	{'precision': 0.7931034482758621, 'recall': 0.06497175141242938, 'f1-score': 0.12010443864229765, 'support': 708.0}	{'precision': 0.35688405797101447, 'recall': 0.09690113133300542, 'f1-score': 0.15241779497098645, 'support': 4066.0}	{'precision': 0.4854771784232365, 'recall': 0.3482142857142857, 'f1-score': 0.4055459272097054, 'support': 2016.0}	{'precision': 0.7254034519284691, 'recall': 0.9546535465354653, 'f1-score': 0.8243874805268375, 'support': 12195.0}	{'precision': 0.8224254998113919, 'recall': 0.8852908334179271, 'f1-score': 0.8527010510877536, 'support': 9851.0}	0.7349	{'precision': 0.4547562337728534, 'recall': 0.3357187926304447, 'f1-score': 0.3364509560625115, 'support': 29261.0}	{'precision': 0.6814305221181916, 'recall': 0.7349372885410614, 'f1-score': 0.6826734788782265, 'support': 29261.0}
No log	2.0	82	0.4757	{'precision': 1.0, 'recall': 0.01056338028169014, 'f1-score': 0.020905923344947737, 'support': 284.0}	{'precision': 0.0, 'recall': 0.0, 'f1-score': 0.0, 'support': 141.0}	{'precision': 0.6255364806866953, 'recall': 0.8234463276836158, 'f1-score': 0.7109756097560975, 'support': 708.0}	{'precision': 0.5658734764944864, 'recall': 0.4795868175110674, 'f1-score': 0.5191693290734825, 'support': 4066.0}	{'precision': 0.745417515274949, 'recall': 0.5446428571428571, 'f1-score': 0.6294067067927773, 'support': 2016.0}	{'precision': 0.8514935768456895, 'recall': 0.9022550225502255, 'f1-score': 0.8761396663614285, 'support': 12195.0}	{'precision': 0.9034811635670005, 'recall': 0.9616282610902447, 'f1-score': 0.931648308418568, 'support': 9851.0}	0.8240	{'precision': 0.6702574589812601, 'recall': 0.5317318094656714, 'f1-score': 0.5268922205353288, 'support': 29261.0}	{'precision': 0.8138696629123667, 'recall': 0.8239636376063703, 'f1-score': 0.8117058591419718, 'support': 29261.0}
No log	3.0	123	0.4101	{'precision': 0.49624060150375937, 'recall': 0.2323943661971831, 'f1-score': 0.31654676258992803, 'support': 284.0}	{'precision': 1.0, 'recall': 0.014184397163120567, 'f1-score': 0.027972027972027972, 'support': 141.0}	{'precision': 0.6877777777777778, 'recall': 0.8742937853107344, 'f1-score': 0.7699004975124378, 'support': 708.0}	{'precision': 0.6374125874125874, 'recall': 0.4483521888834235, 'f1-score': 0.5264221773029165, 'support': 4066.0}	{'precision': 0.7599795291709315, 'recall': 0.7366071428571429, 'f1-score': 0.7481108312342569, 'support': 2016.0}	{'precision': 0.843370836090889, 'recall': 0.9404674046740468, 'f1-score': 0.8892765759478949, 'support': 12195.0}	{'precision': 0.9602568022011617, 'recall': 0.9565526342503299, 'f1-score': 0.9584011391375101, 'support': 9851.0}	0.8505	{'precision': 0.7692911620224437, 'recall': 0.6004074170479973, 'f1-score': 0.6052328588138531, 'support': 29261.0}	{'precision': 0.841977868607838, 'recall': 0.8505177540070401, 'f1-score': 0.8398036418020065, 'support': 29261.0}
No log	4.0	164	0.3859	{'precision': 0.538135593220339, 'recall': 0.4471830985915493, 'f1-score': 0.48846153846153845, 'support': 284.0}	{'precision': 1.0, 'recall': 0.10638297872340426, 'f1-score': 0.19230769230769232, 'support': 141.0}	{'precision': 0.7128146453089245, 'recall': 0.8799435028248588, 'f1-score': 0.7876106194690266, 'support': 708.0}	{'precision': 0.6014307613694431, 'recall': 0.5789473684210527, 'f1-score': 0.5899749373433584, 'support': 4066.0}	{'precision': 0.7848036715961244, 'recall': 0.7633928571428571, 'f1-score': 0.7739502137289415, 'support': 2016.0}	{'precision': 0.8792672100718263, 'recall': 0.8933989339893399, 'f1-score': 0.8862767428617913, 'support': 12195.0}	{'precision': 0.9612968591691996, 'recall': 0.9631509491422191, 'f1-score': 0.9622230110034988, 'support': 9851.0}	0.8558	{'precision': 0.7825355343908367, 'recall': 0.6617713841193258, 'f1-score': 0.6686863935965496, 'support': 29261.0}	{'precision': 0.8550112416363399, 'recall': 0.855780732032398, 'f1-score': 0.8533403597564397, 'support': 29261.0}
No log	5.0	205	0.3818	{'precision': 0.5588235294117647, 'recall': 0.46830985915492956, 'f1-score': 0.5095785440613027, 'support': 284.0}	{'precision': 0.8787878787878788, 'recall': 0.20567375886524822, 'f1-score': 0.3333333333333333, 'support': 141.0}	{'precision': 0.7287735849056604, 'recall': 0.8728813559322034, 'f1-score': 0.794344473007712, 'support': 708.0}	{'precision': 0.6021926389976507, 'recall': 0.5673880964092474, 'f1-score': 0.5842725085475498, 'support': 4066.0}	{'precision': 0.7885196374622356, 'recall': 0.7767857142857143, 'f1-score': 0.782608695652174, 'support': 2016.0}	{'precision': 0.8760707709550877, 'recall': 0.8973349733497334, 'f1-score': 0.8865753868589484, 'support': 12195.0}	{'precision': 0.9648159446817165, 'recall': 0.9631509491422191, 'f1-score': 0.9639827279654559, 'support': 9851.0}	0.8573	{'precision': 0.7711405693145706, 'recall': 0.6787892438770422, 'f1-score': 0.693527952775211, 'support': 29261.0}	{'precision': 0.8552285449410628, 'recall': 0.8572502648576603, 'f1-score': 0.8549088561404111, 'support': 29261.0}

Framework versions

• Transformers 4.45.2
• Pytorch 2.5.0+cu124
• Datasets 2.19.1
• Tokenizers 0.20.1