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-simple
    results:
      - task:
          name: Token Classification
          type: token-classification
        dataset:
          name: essays_su_g
          type: essays_su_g
          config: simple
          split: train[60%:80%]
          args: simple
        metrics:
          - name: Accuracy
            type: accuracy
            value: 0.858776119402985

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: 0.6472
Claim: {'precision': 0.6572622779519331, 'recall': 0.6366396761133604, 'f1-score': 0.6467866323907456, 'support': 4940.0}
Majorclaim: {'precision': 0.8274678111587983, 'recall': 0.8811700182815356, 'f1-score': 0.8534749889331562, 'support': 2188.0}
O: {'precision': 0.9268028016178357, 'recall': 0.8970686527260575, 'f1-score': 0.9116933527413877, 'support': 10473.0}
Premise: {'precision': 0.8801698670605613, 'recall': 0.8994905339958488, 'f1-score': 0.8897253242915357, 'support': 15899.0}
Accuracy: 0.8588
Macro avg: {'precision': 0.8229256894472821, 'recall': 0.8285922202792007, 'f1-score': 0.8254200745892063, 'support': 33500.0}
Weighted avg: {'precision': 0.8584358710936555, 'recall': 0.858776119402985, 'f1-score': 0.8584010941482899, 'support': 33500.0}

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: 16

Training results

Training Loss	Epoch	Step	Validation Loss	Claim	Majorclaim	O	Premise	Accuracy	Macro avg	Weighted avg
No log	1.0	41	0.6237	{'precision': 0.4813399941228328, 'recall': 0.33157894736842103, 'f1-score': 0.3926645091693635, 'support': 4940.0}	{'precision': 0.41758530183727033, 'recall': 0.7271480804387569, 'f1-score': 0.5305101700566855, 'support': 2188.0}	{'precision': 0.8614998552263295, 'recall': 0.8522868328081734, 'f1-score': 0.8568685802054334, 'support': 10473.0}	{'precision': 0.8528192892126083, 'recall': 0.8542675639977357, 'f1-score': 0.8535428122545169, 'support': 15899.0}	0.7683	{'precision': 0.6533111100997602, 'recall': 0.6913203561532717, 'f1-score': 0.6583965179214999, 'support': 33500.0}	{'precision': 0.7723271066974134, 'recall': 0.7682686567164179, 'f1-score': 0.7655218131315448, 'support': 33500.0}
No log	2.0	82	0.4751	{'precision': 0.5846230654018971, 'recall': 0.47408906882591095, 'f1-score': 0.5235859602056785, 'support': 4940.0}	{'precision': 0.7269767441860465, 'recall': 0.7143510054844607, 'f1-score': 0.7206085753803596, 'support': 2188.0}	{'precision': 0.9142337609859582, 'recall': 0.8641268022534135, 'f1-score': 0.8884743765953269, 'support': 10473.0}	{'precision': 0.8357695614789338, 'recall': 0.917038807472168, 'f1-score': 0.8745201535508637, 'support': 15899.0}	0.8219	{'precision': 0.7654007830132088, 'recall': 0.7424014210089883, 'f1-score': 0.7517972664330571, 'support': 33500.0}	{'precision': 0.816159208839521, 'recall': 0.8219402985074626, 'f1-score': 0.8170804260816812, 'support': 33500.0}
No log	3.0	123	0.4586	{'precision': 0.6658894070619586, 'recall': 0.4046558704453441, 'f1-score': 0.5033996474439688, 'support': 4940.0}	{'precision': 0.7872244714349977, 'recall': 0.79981718464351, 'f1-score': 0.7934708682838358, 'support': 2188.0}	{'precision': 0.9342819121711536, 'recall': 0.8714790413444095, 'f1-score': 0.9017883608339096, 'support': 10473.0}	{'precision': 0.8168702042580784, 'recall': 0.9508145166362665, 'f1-score': 0.8787676209853219, 'support': 15899.0}	0.8356	{'precision': 0.8010664987315472, 'recall': 0.7566916532673825, 'f1-score': 0.7693566243867591, 'support': 33500.0}	{'precision': 0.8293759599418965, 'recall': 0.8356119402985075, 'f1-score': 0.8250407291712659, 'support': 33500.0}
No log	4.0	164	0.4525	{'precision': 0.5575898801597869, 'recall': 0.6781376518218624, 'f1-score': 0.6119839240043845, 'support': 4940.0}	{'precision': 0.7466456195737964, 'recall': 0.8647166361974405, 'f1-score': 0.8013553578991952, 'support': 2188.0}	{'precision': 0.9201592832254853, 'recall': 0.8825551417931825, 'f1-score': 0.9009650063359004, 'support': 10473.0}	{'precision': 0.8922416683430564, 'recall': 0.836907981634065, 'f1-score': 0.8636894716344281, 'support': 15899.0}	0.8296	{'precision': 0.7791591128255312, 'recall': 0.8155793528616375, 'f1-score': 0.7944984399684771, 'support': 33500.0}	{'precision': 0.8421114352783158, 'recall': 0.8295820895522388, 'f1-score': 0.8341543739861718, 'support': 33500.0}
No log	5.0	205	0.4721	{'precision': 0.662877030162413, 'recall': 0.5783400809716599, 'f1-score': 0.6177297297297297, 'support': 4940.0}	{'precision': 0.7945205479452054, 'recall': 0.8747714808043876, 'f1-score': 0.8327169893408746, 'support': 2188.0}	{'precision': 0.9125229313507772, 'recall': 0.9024157357013273, 'f1-score': 0.9074411905904946, 'support': 10473.0}	{'precision': 0.8726254262055528, 'recall': 0.9014403421598842, 'f1-score': 0.8867988738669058, 'support': 15899.0}	0.8524	{'precision': 0.8106364839159872, 'recall': 0.8142419099093148, 'f1-score': 0.8111716958820011, 'support': 33500.0}	{'precision': 0.8490670984831405, 'recall': 0.8523582089552239, 'f1-score': 0.8500422842449816, 'support': 33500.0}
No log	6.0	246	0.4792	{'precision': 0.6428419936373276, 'recall': 0.6135627530364373, 'f1-score': 0.6278612118073537, 'support': 4940.0}	{'precision': 0.804950917626974, 'recall': 0.8619744058500914, 'f1-score': 0.83248730964467, 'support': 2188.0}	{'precision': 0.9285714285714286, 'recall': 0.8949680129857729, 'f1-score': 0.9114601059950406, 'support': 10473.0}	{'precision': 0.872155615365794, 'recall': 0.8967859613812189, 'f1-score': 0.8842993146649301, 'support': 15899.0}	0.8522	{'precision': 0.812129988800381, 'recall': 0.8168227833133802, 'f1-score': 0.8140269855279986, 'support': 33500.0}	{'precision': 0.8515881419840463, 'recall': 0.8521791044776119, 'f1-score': 0.8515914362320791, 'support': 33500.0}
No log	7.0	287	0.5202	{'precision': 0.6744186046511628, 'recall': 0.5342105263157895, 'f1-score': 0.5961820851688694, 'support': 4940.0}	{'precision': 0.8121475054229935, 'recall': 0.8555758683729433, 'f1-score': 0.8332962385933673, 'support': 2188.0}	{'precision': 0.9198786930150655, 'recall': 0.8978325217225246, 'f1-score': 0.9087219135056778, 'support': 10473.0}	{'precision': 0.8582063305978898, 'recall': 0.9208755267626895, 'f1-score': 0.8884371491853515, 'support': 15899.0}	0.8524	{'precision': 0.816162783421778, 'recall': 0.8021236107934867, 'f1-score': 0.8066593466133165, 'support': 33500.0}	{'precision': 0.8473766761482054, 'recall': 0.8523880597014926, 'f1-score': 0.8480805524125185, 'support': 33500.0}
No log	8.0	328	0.5458	{'precision': 0.6705622932745314, 'recall': 0.6155870445344129, 'f1-score': 0.6418997361477573, 'support': 4940.0}	{'precision': 0.8129251700680272, 'recall': 0.8738574040219378, 'f1-score': 0.8422907488986784, 'support': 2188.0}	{'precision': 0.9259259259259259, 'recall': 0.89277188962093, 'f1-score': 0.909046716251033, 'support': 10473.0}	{'precision': 0.8728428701180745, 'recall': 0.9066607962764954, 'f1-score': 0.8894304929968533, 'support': 15899.0}	0.8573	{'precision': 0.8205640648466398, 'recall': 0.822219283613444, 'f1-score': 0.8206669235735804, 'support': 33500.0}	{'precision': 0.8556957914959558, 'recall': 0.8572537313432835, 'f1-score': 0.8559826424660975, 'support': 33500.0}
No log	9.0	369	0.5550	{'precision': 0.6423661737138097, 'recall': 0.6242914979757085, 'f1-score': 0.6331998768093625, 'support': 4940.0}	{'precision': 0.8291592128801432, 'recall': 0.8473491773308958, 'f1-score': 0.8381555153707052, 'support': 2188.0}	{'precision': 0.909720885466795, 'recall': 0.9025112193258856, 'f1-score': 0.9061017111633034, 'support': 10473.0}	{'precision': 0.8796739874323399, 'recall': 0.8893012139128247, 'f1-score': 0.8844614037282621, 'support': 15899.0}	0.8516	{'precision': 0.8152300648732719, 'recall': 0.8158632771363286, 'f1-score': 0.8154796267679083, 'support': 33500.0}	{'precision': 0.8507741138987609, 'recall': 0.8516119402985075, 'f1-score': 0.8511506488942767, 'support': 33500.0}
No log	10.0	410	0.5788	{'precision': 0.6611198560827524, 'recall': 0.5951417004048583, 'f1-score': 0.6263982102908278, 'support': 4940.0}	{'precision': 0.8315460232350312, 'recall': 0.8505484460694699, 'f1-score': 0.8409399005874378, 'support': 2188.0}	{'precision': 0.9248446592366111, 'recall': 0.8953499474840065, 'f1-score': 0.9098583349505143, 'support': 10473.0}	{'precision': 0.8645358599184456, 'recall': 0.9067865903515945, 'f1-score': 0.8851573292402148, 'support': 15899.0}	0.8536	{'precision': 0.8205115996182102, 'recall': 0.8119566710774824, 'f1-score': 0.8155884437672487, 'support': 33500.0}	{'precision': 0.851239060922849, 'recall': 0.8535820895522388, 'f1-score': 0.8518342203238483, 'support': 33500.0}
No log	11.0	451	0.5865	{'precision': 0.661878453038674, 'recall': 0.6062753036437247, 'f1-score': 0.6328578975171685, 'support': 4940.0}	{'precision': 0.829535495179667, 'recall': 0.8651736745886655, 'f1-score': 0.8469798657718122, 'support': 2188.0}	{'precision': 0.9291244788564622, 'recall': 0.8937267258665139, 'f1-score': 0.9110819097678493, 'support': 10473.0}	{'precision': 0.8703893134364282, 'recall': 0.9098056481539719, 'f1-score': 0.88966111076942, 'support': 15899.0}	0.8571	{'precision': 0.8227319351278078, 'recall': 0.818745338063219, 'f1-score': 0.8201451959565625, 'support': 33500.0}	{'precision': 0.8553356293389153, 'recall': 0.8571044776119403, 'f1-score': 0.8557012776467233, 'support': 33500.0}
No log	12.0	492	0.6140	{'precision': 0.6268885064065787, 'recall': 0.6635627530364372, 'f1-score': 0.6447044940505456, 'support': 4940.0}	{'precision': 0.8325078793336335, 'recall': 0.8450639853747715, 'f1-score': 0.8387389430709912, 'support': 2188.0}	{'precision': 0.923546196989078, 'recall': 0.896209300105032, 'f1-score': 0.9096724171351037, 'support': 10473.0}	{'precision': 0.885440926543715, 'recall': 0.8847726272092584, 'f1-score': 0.885106650726735, 'support': 15899.0}	0.8531	{'precision': 0.8170958773182513, 'recall': 0.8224021664313748, 'f1-score': 0.8195556262458439, 'support': 33500.0}	{'precision': 0.8557695842930038, 'recall': 0.8531343283582089, 'f1-score': 0.8543077872420695, 'support': 33500.0}
0.2701	13.0	533	0.6368	{'precision': 0.6831773567678612, 'recall': 0.6058704453441296, 'f1-score': 0.642205771912885, 'support': 4940.0}	{'precision': 0.8174536256323778, 'recall': 0.8861974405850092, 'f1-score': 0.8504385964912281, 'support': 2188.0}	{'precision': 0.9274289099526066, 'recall': 0.8968776854769407, 'f1-score': 0.9118974807048201, 'support': 10473.0}	{'precision': 0.8733377459534268, 'recall': 0.912887602993899, 'f1-score': 0.892674826250077, 'support': 15899.0}	0.8609	{'precision': 0.8253494095765681, 'recall': 0.8254582935999946, 'f1-score': 0.8243041688397525, 'support': 33500.0}	{'precision': 0.8585565514078823, 'recall': 0.8608656716417911, 'f1-score': 0.8589909116520602, 'support': 33500.0}
0.2701	14.0	574	0.6486	{'precision': 0.6641386782231853, 'recall': 0.6204453441295547, 'f1-score': 0.641548927263213, 'support': 4940.0}	{'precision': 0.8142076502732241, 'recall': 0.8852833638025595, 'f1-score': 0.8482592511495511, 'support': 2188.0}	{'precision': 0.9240070782540307, 'recall': 0.897450587224291, 'f1-score': 0.9105352385565513, 'support': 10473.0}	{'precision': 0.8767601322395004, 'recall': 0.9007484747468394, 'f1-score': 0.888592436323023, 'support': 15899.0}	0.8574	{'precision': 0.8197783847474851, 'recall': 0.8259819424758111, 'f1-score': 0.8222339633230846, 'support': 33500.0}	{'precision': 0.8560915487238994, 'recall': 0.8573731343283582, 'f1-score': 0.8563883474835222, 'support': 33500.0}
0.2701	15.0	615	0.6462	{'precision': 0.6603214890016921, 'recall': 0.6319838056680162, 'f1-score': 0.6458419528340918, 'support': 4940.0}	{'precision': 0.8342832091188075, 'recall': 0.8697440585009141, 'f1-score': 0.8516446632356232, 'support': 2188.0}	{'precision': 0.9237646134197859, 'recall': 0.8978325217225246, 'f1-score': 0.9106139841177611, 'support': 10473.0}	{'precision': 0.8785556645414418, 'recall': 0.9013774451223348, 'f1-score': 0.8898202477414549, 'support': 15899.0}	0.8585	{'precision': 0.8242312440204318, 'recall': 0.8252344577534474, 'f1-score': 0.8244802119822328, 'support': 33500.0}	{'precision': 0.8576162126600033, 'recall': 0.8584776119402985, 'f1-score': 0.8578498550646764, 'support': 33500.0}
0.2701	16.0	656	0.6472	{'precision': 0.6572622779519331, 'recall': 0.6366396761133604, 'f1-score': 0.6467866323907456, 'support': 4940.0}	{'precision': 0.8274678111587983, 'recall': 0.8811700182815356, 'f1-score': 0.8534749889331562, 'support': 2188.0}	{'precision': 0.9268028016178357, 'recall': 0.8970686527260575, 'f1-score': 0.9116933527413877, 'support': 10473.0}	{'precision': 0.8801698670605613, 'recall': 0.8994905339958488, 'f1-score': 0.8897253242915357, 'support': 15899.0}	0.8588	{'precision': 0.8229256894472821, 'recall': 0.8285922202792007, 'f1-score': 0.8254200745892063, 'support': 33500.0}	{'precision': 0.8584358710936555, 'recall': 0.858776119402985, 'f1-score': 0.8584010941482899, 'support': 33500.0}

Framework versions

Transformers 4.37.2
Pytorch 2.2.0+cu121
Datasets 2.17.0
Tokenizers 0.15.2