metadata

license: mit
base_model: SCUT-DLVCLab/lilt-roberta-en-base
tags:
  - generated_from_trainer
model-index:
  - name: lilt-en-funsd
    results: []

lilt-en-funsd

This model is a fine-tuned version of SCUT-DLVCLab/lilt-roberta-en-base on an unknown dataset. It achieves the following results on the evaluation set:

Loss: 1.6214
Answer: {'precision': 0.8792899408284024, 'recall': 0.9094247246022031, 'f1': 0.8941034897713599, 'number': 817}
Header: {'precision': 0.6078431372549019, 'recall': 0.5210084033613446, 'f1': 0.5610859728506787, 'number': 119}
Question: {'precision': 0.8960573476702509, 'recall': 0.9285051067780873, 'f1': 0.9119927040583674, 'number': 1077}
Overall Precision: 0.8749
Overall Recall: 0.8967
Overall F1: 0.8857
Overall Accuracy: 0.8129

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: 5e-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
training_steps: 2500
mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Answer	Header	Question	Overall Precision	Overall Recall	Overall F1	Overall Accuracy
0.4022	10.5263	200	1.0717	{'precision': 0.8156424581005587, 'recall': 0.8935128518971848, 'f1': 0.852803738317757, 'number': 817}	{'precision': 0.452991452991453, 'recall': 0.44537815126050423, 'f1': 0.4491525423728814, 'number': 119}	{'precision': 0.8516579406631762, 'recall': 0.9062209842154132, 'f1': 0.8780926675663517, 'number': 1077}	0.8151	0.8738	0.8434	0.7966
0.0484	21.0526	400	1.2460	{'precision': 0.8537794299876085, 'recall': 0.8433292533659731, 'f1': 0.8485221674876847, 'number': 817}	{'precision': 0.5454545454545454, 'recall': 0.5546218487394958, 'f1': 0.5499999999999999, 'number': 119}	{'precision': 0.8713398402839396, 'recall': 0.9117920148560817, 'f1': 0.8911070780399274, 'number': 1077}	0.8453	0.8629	0.8540	0.8008
0.0143	31.5789	600	1.5585	{'precision': 0.8566433566433567, 'recall': 0.8996328029375765, 'f1': 0.8776119402985075, 'number': 817}	{'precision': 0.5294117647058824, 'recall': 0.5294117647058824, 'f1': 0.5294117647058824, 'number': 119}	{'precision': 0.8879310344827587, 'recall': 0.8607242339832869, 'f1': 0.8741159830268741, 'number': 1077}	0.8535	0.8569	0.8552	0.7935
0.0079	42.1053	800	1.5146	{'precision': 0.8556338028169014, 'recall': 0.8922888616891065, 'f1': 0.8735769922109047, 'number': 817}	{'precision': 0.47761194029850745, 'recall': 0.5378151260504201, 'f1': 0.5059288537549407, 'number': 119}	{'precision': 0.8851540616246498, 'recall': 0.8802228412256268, 'f1': 0.88268156424581, 'number': 1077}	0.8464	0.8649	0.8555	0.7989
0.0041	52.6316	1000	1.5279	{'precision': 0.8536299765807962, 'recall': 0.8922888616891065, 'f1': 0.8725314183123878, 'number': 817}	{'precision': 0.6, 'recall': 0.5042016806722689, 'f1': 0.547945205479452, 'number': 119}	{'precision': 0.8764342453662842, 'recall': 0.9220055710306406, 'f1': 0.8986425339366516, 'number': 1077}	0.8539	0.8852	0.8693	0.8063
0.0031	63.1579	1200	1.5225	{'precision': 0.8413793103448276, 'recall': 0.8959608323133414, 'f1': 0.8678126852400712, 'number': 817}	{'precision': 0.5794392523364486, 'recall': 0.5210084033613446, 'f1': 0.5486725663716815, 'number': 119}	{'precision': 0.8873499538319483, 'recall': 0.8922934076137419, 'f1': 0.8898148148148148, 'number': 1077}	0.8519	0.8718	0.8618	0.8014
0.0016	73.6842	1400	1.6214	{'precision': 0.8792899408284024, 'recall': 0.9094247246022031, 'f1': 0.8941034897713599, 'number': 817}	{'precision': 0.6078431372549019, 'recall': 0.5210084033613446, 'f1': 0.5610859728506787, 'number': 119}	{'precision': 0.8960573476702509, 'recall': 0.9285051067780873, 'f1': 0.9119927040583674, 'number': 1077}	0.8749	0.8967	0.8857	0.8129
0.0011	84.2105	1600	1.7158	{'precision': 0.8770186335403727, 'recall': 0.8641370869033048, 'f1': 0.8705302096177557, 'number': 817}	{'precision': 0.5419847328244275, 'recall': 0.5966386554621849, 'f1': 0.568, 'number': 119}	{'precision': 0.9052044609665427, 'recall': 0.904363974001857, 'f1': 0.9047840222944729, 'number': 1077}	0.8703	0.8698	0.8701	0.8070
0.001	94.7368	1800	1.6261	{'precision': 0.8481735159817352, 'recall': 0.9094247246022031, 'f1': 0.8777318369757826, 'number': 817}	{'precision': 0.6017699115044248, 'recall': 0.5714285714285714, 'f1': 0.5862068965517241, 'number': 119}	{'precision': 0.9050751879699248, 'recall': 0.8941504178272981, 'f1': 0.8995796356842597, 'number': 1077}	0.8641	0.8813	0.8726	0.8128
0.0004	105.2632	2000	1.6253	{'precision': 0.8611435239206534, 'recall': 0.9033047735618115, 'f1': 0.8817204301075269, 'number': 817}	{'precision': 0.625, 'recall': 0.5462184873949579, 'f1': 0.5829596412556054, 'number': 119}	{'precision': 0.894404332129964, 'recall': 0.9201485608170845, 'f1': 0.9070938215102976, 'number': 1077}	0.8671	0.8912	0.8790	0.8194
0.0001	115.7895	2200	1.6711	{'precision': 0.8823529411764706, 'recall': 0.8812729498164015, 'f1': 0.8818126148193509, 'number': 817}	{'precision': 0.6283185840707964, 'recall': 0.5966386554621849, 'f1': 0.6120689655172413, 'number': 119}	{'precision': 0.8967391304347826, 'recall': 0.9192200557103064, 'f1': 0.9078404401650619, 'number': 1077}	0.8760	0.8847	0.8804	0.8190
0.0002	126.3158	2400	1.6682	{'precision': 0.8756038647342995, 'recall': 0.8873929008567931, 'f1': 0.8814589665653496, 'number': 817}	{'precision': 0.6330275229357798, 'recall': 0.5798319327731093, 'f1': 0.6052631578947367, 'number': 119}	{'precision': 0.8997289972899729, 'recall': 0.924791086350975, 'f1': 0.9120879120879122, 'number': 1077}	0.8757	0.8892	0.8824	0.8179

Framework versions

Transformers 4.40.2
Pytorch 2.2.1+cu121
Datasets 2.19.1
Tokenizers 0.19.1