meta_data/README_s42_e5.md

metadata

license: apache-2.0
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: test
          args: spans
        metrics:
          - name: Accuracy
            type: accuracy
            value: 0.9400193485972267

longformer-spans

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: 0.1716
• B: {'precision': 0.8278829604130808, 'recall': 0.9084041548630784, 'f1-score': 0.8662764520486267, 'support': 1059.0}
• I: {'precision': 0.949054915557544, 'recall': 0.9656330014224751, 'f1-score': 0.9572721888484643, 'support': 17575.0}
• O: {'precision': 0.9364918217710095, 'recall': 0.8950943396226415, 'f1-score': 0.9153252480705623, 'support': 9275.0}
• Accuracy: 0.9400
• Macro avg: {'precision': 0.9044765659138781, 'recall': 0.9230438319693982, 'f1-score': 0.9129579629892177, 'support': 27909.0}
• Weighted avg: {'precision': 0.9402819822611845, 'recall': 0.9400193485972267, 'f1-score': 0.9398791485752167, 'support': 27909.0}

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	I	O	Accuracy	Macro avg	Weighted avg
No log	1.0	41	0.2820	{'precision': 0.8252595155709342, 'recall': 0.45042492917847027, 'f1-score': 0.582773365913256, 'support': 1059.0}	{'precision': 0.9113681210260908, 'recall': 0.9460597439544808, 'f1-score': 0.9283899606354169, 'support': 17575.0}	{'precision': 0.879608231539562, 'recall': 0.8617789757412398, 'f1-score': 0.8706023309007732, 'support': 9275.0}	0.8992	{'precision': 0.8720786227121957, 'recall': 0.7527545496247302, 'f1-score': 0.793921885816482, 'support': 27909.0}	{'precision': 0.8975459852217064, 'recall': 0.8992439714787345, 'f1-score': 0.896071058503503, 'support': 27909.0}
No log	2.0	82	0.1953	{'precision': 0.812897366030881, 'recall': 0.8451369216241738, 'f1-score': 0.8287037037037038, 'support': 1059.0}	{'precision': 0.9452124358178637, 'recall': 0.9531721194879089, 'f1-score': 0.9491755906850247, 'support': 17575.0}	{'precision': 0.9121629058888278, 'recall': 0.8934770889487871, 'f1-score': 0.902723311546841, 'support': 9275.0}	0.9292	{'precision': 0.8900909025791909, 'recall': 0.8972620433536234, 'f1-score': 0.8935342019785232, 'support': 27909.0}	{'precision': 0.9292084210199053, 'recall': 0.9292342971801211, 'f1-score': 0.9291668258665118, 'support': 27909.0}
No log	3.0	123	0.1858	{'precision': 0.7883211678832117, 'recall': 0.9178470254957507, 'f1-score': 0.8481675392670156, 'support': 1059.0}	{'precision': 0.9373831775700935, 'recall': 0.9701849217638692, 'f1-score': 0.9535020271214875, 'support': 17575.0}	{'precision': 0.9481498939429649, 'recall': 0.8674932614555256, 'f1-score': 0.9060300658746692, 'support': 9275.0}	0.9341	{'precision': 0.8912847464654234, 'recall': 0.9185084029050485, 'f1-score': 0.9025665440877241, 'support': 27909.0}	{'precision': 0.9353051606615684, 'recall': 0.9340714464867964, 'f1-score': 0.9337287760841115, 'support': 27909.0}
No log	4.0	164	0.1704	{'precision': 0.8296943231441049, 'recall': 0.8970727101038716, 'f1-score': 0.8620689655172413, 'support': 1059.0}	{'precision': 0.9604448520981427, 'recall': 0.9532859174964438, 'f1-score': 0.9568519946314857, 'support': 17575.0}	{'precision': 0.9158798283261803, 'recall': 0.9203234501347709, 'f1-score': 0.9180962624361388, 'support': 9275.0}	0.9402	{'precision': 0.9020063345228092, 'recall': 0.923560692578362, 'f1-score': 0.9123390741949553, 'support': 27909.0}	{'precision': 0.9406732585029842, 'recall': 0.9401985022752517, 'f1-score': 0.9403757810823142, 'support': 27909.0}
No log	5.0	205	0.1716	{'precision': 0.8278829604130808, 'recall': 0.9084041548630784, 'f1-score': 0.8662764520486267, 'support': 1059.0}	{'precision': 0.949054915557544, 'recall': 0.9656330014224751, 'f1-score': 0.9572721888484643, 'support': 17575.0}	{'precision': 0.9364918217710095, 'recall': 0.8950943396226415, 'f1-score': 0.9153252480705623, 'support': 9275.0}	0.9400	{'precision': 0.9044765659138781, 'recall': 0.9230438319693982, 'f1-score': 0.9129579629892177, 'support': 27909.0}	{'precision': 0.9402819822611845, 'recall': 0.9400193485972267, 'f1-score': 0.9398791485752167, 'support': 27909.0}

Framework versions

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