meta_data/README_s42_e9.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-sep_tok
    results:
      - task:
          name: Token Classification
          type: token-classification
        dataset:
          name: essays_su_g
          type: essays_su_g
          config: sep_tok
          split: train[80%:100%]
          args: sep_tok
        metrics:
          - name: Accuracy
            type: accuracy
            value: 0.8961117656960108

longformer-sep_tok

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: 0.3184
• Claim: {'precision': 0.6251433815095205, 'recall': 0.6537907869481766, 'f1-score': 0.6391462413510027, 'support': 4168.0}
• Majorclaim: {'precision': 0.8963831867057673, 'recall': 0.8522304832713755, 'f1-score': 0.873749404478323, 'support': 2152.0}
• O: {'precision': 1.0, 'recall': 0.999557991513437, 'f1-score': 0.9997789469030461, 'support': 11312.0}
• Premise: {'precision': 0.8966063537063287, 'recall': 0.8906651205168558, 'f1-score': 0.8936258622122497, 'support': 12073.0}
• Accuracy: 0.8961
• Macro avg: {'precision': 0.8545332304804041, 'recall': 0.8490610955624612, 'f1-score': 0.8515751137361555, 'support': 29705.0}
• Weighted avg: {'precision': 0.8978738508742299, 'recall': 0.8961117656960108, 'f1-score': 0.8969033743223053, 'support': 29705.0}

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: 9

Training results

Training Loss	Epoch	Step	Validation Loss	Claim	Majorclaim	O	Premise	Accuracy	Macro avg	Weighted avg
No log	1.0	41	0.3777	{'precision': 0.44112006137322596, 'recall': 0.2759117082533589, 'f1-score': 0.33948339483394835, 'support': 4168.0}	{'precision': 0.7229235880398671, 'recall': 0.5055762081784386, 'f1-score': 0.5950232430954334, 'support': 2152.0}	{'precision': 0.9991005576542543, 'recall': 0.981966053748232, 'f1-score': 0.9904592064199733, 'support': 11312.0}	{'precision': 0.7974438687392055, 'recall': 0.9561003892984345, 'f1-score': 0.8695946963989755, 'support': 12073.0}	0.8379	{'precision': 0.7401470189516381, 'recall': 0.6798885898696161, 'f1-score': 0.6986401351870827, 'support': 29705.0}	{'precision': 0.8188414513630281, 'recall': 0.8378724120518432, 'f1-score': 0.8213481946290806, 'support': 29705.0}
No log	2.0	82	0.2665	{'precision': 0.6216867469879518, 'recall': 0.3714011516314779, 'f1-score': 0.4650045058576149, 'support': 4168.0}	{'precision': 0.7490087232355274, 'recall': 0.8777881040892194, 'f1-score': 0.808301240907146, 'support': 2152.0}	{'precision': 0.9999115044247787, 'recall': 0.9988507779349364, 'f1-score': 0.9993808597205024, 'support': 11312.0}	{'precision': 0.8468603001567984, 'recall': 0.9394516690135012, 'f1-score': 0.8907563025210085, 'support': 12073.0}	0.8779	{'precision': 0.8043668187012641, 'recall': 0.7968729256672837, 'f1-score': 0.7908607272515679, 'support': 29705.0}	{'precision': 0.8664602079008504, 'recall': 0.8778993435448578, 'f1-score': 0.8664097012738992, 'support': 29705.0}
No log	3.0	123	0.2423	{'precision': 0.6248701973001038, 'recall': 0.5774952015355086, 'f1-score': 0.6002493765586034, 'support': 4168.0}	{'precision': 0.8225806451612904, 'recall': 0.8531598513011153, 'f1-score': 0.8375912408759125, 'support': 2152.0}	{'precision': 1.0, 'recall': 0.9994695898161244, 'f1-score': 0.9997347245556637, 'support': 11312.0}	{'precision': 0.886398700771417, 'recall': 0.904166321543941, 'f1-score': 0.895194357880925, 'support': 12073.0}	0.8909	{'precision': 0.8334623858082029, 'recall': 0.8335727410491722, 'f1-score': 0.8331924249677761, 'support': 29705.0}	{'precision': 0.8883401462766283, 'recall': 0.8909274532906918, 'f1-score': 0.8894467745743578, 'support': 29705.0}
No log	4.0	164	0.2482	{'precision': 0.6370126304228446, 'recall': 0.5566218809980806, 'f1-score': 0.5941101152368758, 'support': 4168.0}	{'precision': 0.85006753714543, 'recall': 0.8773234200743495, 'f1-score': 0.8634804482048937, 'support': 2152.0}	{'precision': 1.0, 'recall': 0.9997347949080623, 'f1-score': 0.9998673798682641, 'support': 11312.0}	{'precision': 0.8778424958110588, 'recall': 0.9112896546011762, 'f1-score': 0.8942534341217588, 'support': 12073.0}	0.8927	{'precision': 0.8412306658448333, 'recall': 0.8362424376454172, 'f1-score': 0.8379278443579481, 'support': 29705.0}	{'precision': 0.8885576985512976, 'recall': 0.8927453290691802, 'f1-score': 0.890129015184852, 'support': 29705.0}
No log	5.0	205	0.2617	{'precision': 0.6225581395348837, 'recall': 0.642274472168906, 'f1-score': 0.6322626358053851, 'support': 4168.0}	{'precision': 0.8676403468735737, 'recall': 0.883364312267658, 'f1-score': 0.8754317292194335, 'support': 2152.0}	{'precision': 1.0, 'recall': 0.9994695898161244, 'f1-score': 0.9997347245556637, 'support': 11312.0}	{'precision': 0.8983036614040981, 'recall': 0.8860266710842376, 'f1-score': 0.8921229306534341, 'support': 12073.0}	0.8948	{'precision': 0.8471255369531389, 'recall': 0.8527837613342315, 'f1-score': 0.8498880050584791, 'support': 29705.0}	{'precision': 0.8961186485839084, 'recall': 0.8948325197778152, 'f1-score': 0.8954317149728881, 'support': 29705.0}
No log	6.0	246	0.2822	{'precision': 0.5918902562033488, 'recall': 0.7039347408829175, 'f1-score': 0.6430684931506849, 'support': 4168.0}	{'precision': 0.9098445595854923, 'recall': 0.8159851301115242, 'f1-score': 0.8603625673689368, 'support': 2152.0}	{'precision': 1.0, 'recall': 0.9996463932107497, 'f1-score': 0.9998231653404068, 'support': 11312.0}	{'precision': 0.9072111207645526, 'recall': 0.8649051602749938, 'f1-score': 0.8855531526947378, 'support': 12073.0}	0.8901	{'precision': 0.8522364841383483, 'recall': 0.8461178561200462, 'f1-score': 0.8472018446386915, 'support': 29705.0}	{'precision': 0.8984933156395887, 'recall': 0.8900858441339842, 'f1-score': 0.8932197469531816, 'support': 29705.0}
No log	7.0	287	0.3260	{'precision': 0.6209569633787757, 'recall': 0.557341650671785, 'f1-score': 0.5874320394487292, 'support': 4168.0}	{'precision': 0.9243511871893981, 'recall': 0.7778810408921933, 'f1-score': 0.8448145344436034, 'support': 2152.0}	{'precision': 1.0, 'recall': 0.9997347949080623, 'f1-score': 0.9998673798682641, 'support': 11312.0}	{'precision': 0.8664746184989099, 'recall': 0.9218089952787211, 'f1-score': 0.893285708552394, 'support': 12073.0}	0.8899	{'precision': 0.852945692266771, 'recall': 0.8141916204376904, 'f1-score': 0.8313499155782478, 'support': 29705.0}	{'precision': 0.8870661655388541, 'recall': 0.8899175223026426, 'f1-score': 0.8874464157201748, 'support': 29705.0}
No log	8.0	328	0.3097	{'precision': 0.6349094330745707, 'recall': 0.647552783109405, 'f1-score': 0.6411687848913172, 'support': 4168.0}	{'precision': 0.896469465648855, 'recall': 0.8731412639405205, 'f1-score': 0.8846516007532956, 'support': 2152.0}	{'precision': 1.0, 'recall': 0.9996463932107497, 'f1-score': 0.9998231653404068, 'support': 11312.0}	{'precision': 0.8954356846473029, 'recall': 0.89372981032055, 'f1-score': 0.8945819342536168, 'support': 12073.0}	0.8980	{'precision': 0.8567036458426821, 'recall': 0.8535175626453062, 'f1-score': 0.8550563713096592, 'support': 29705.0}	{'precision': 0.8987746112734568, 'recall': 0.8980306345733041, 'f1-score': 0.8983823961899579, 'support': 29705.0}
No log	9.0	369	0.3184	{'precision': 0.6251433815095205, 'recall': 0.6537907869481766, 'f1-score': 0.6391462413510027, 'support': 4168.0}	{'precision': 0.8963831867057673, 'recall': 0.8522304832713755, 'f1-score': 0.873749404478323, 'support': 2152.0}	{'precision': 1.0, 'recall': 0.999557991513437, 'f1-score': 0.9997789469030461, 'support': 11312.0}	{'precision': 0.8966063537063287, 'recall': 0.8906651205168558, 'f1-score': 0.8936258622122497, 'support': 12073.0}	0.8961	{'precision': 0.8545332304804041, 'recall': 0.8490610955624612, 'f1-score': 0.8515751137361555, 'support': 29705.0}	{'precision': 0.8978738508742299, 'recall': 0.8961117656960108, 'f1-score': 0.8969033743223053, 'support': 29705.0}

Framework versions

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