layoutlmv3-triplet

This model is a fine-tuned version of microsoft/layoutlmv3-base on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 0.0158
• Aption: {'precision': 0.9251446070091868, 'recall': 0.9238871899422358, 'f1': 0.9245154709282557, 'number': 2943}
• Ootnote: {'precision': 0.9455411844792376, 'recall': 0.9442556084296397, 'f1': 0.9448979591836736, 'number': 2942}
• Overall Precision: 0.9353
• Overall Recall: 0.9341
• Overall F1: 0.9347
• Overall Accuracy: 0.9982

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: 3e-05
• train_batch_size: 8
• eval_batch_size: 4
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 10

Training results

Training Loss	Epoch	Step	Validation Loss	Aption	Ootnote	Overall Precision	Overall Recall	Overall F1	Overall Accuracy
0.0152	1.0	8507	0.0147	{'precision': 0.863031914893617, 'recall': 0.8820931022765885, 'f1': 0.8724584103512015, 'number': 2943}	{'precision': 0.902593085106383, 'recall': 0.9228416043507818, 'f1': 0.9126050420168068, 'number': 2942}	0.8828	0.9025	0.8925	0.9969
0.0067	2.0	17014	0.0128	{'precision': 0.9206239168110919, 'recall': 0.9024804621134896, 'f1': 0.911461908030199, 'number': 2943}	{'precision': 0.9459084604715673, 'recall': 0.9272603670972128, 'f1': 0.936491589426708, 'number': 2942}	0.9333	0.9149	0.9240	0.9979
0.0049	3.0	25521	0.0153	{'precision': 0.9005291005291005, 'recall': 0.8674821610601428, 'f1': 0.8836967808930426, 'number': 2943}	{'precision': 0.9435426958362738, 'recall': 0.9089055064581917, 'f1': 0.9259002770083102, 'number': 2942}	0.9220	0.8882	0.9048	0.9971
0.0037	4.0	34028	0.0110	{'precision': 0.9221803222488858, 'recall': 0.9140332993544003, 'f1': 0.9180887372013652, 'number': 2943}	{'precision': 0.946159122085048, 'recall': 0.9377974167233175, 'f1': 0.9419597132127007, 'number': 2942}	0.9342	0.9259	0.9300	0.9981
0.0025	5.0	42535	0.0110	{'precision': 0.9253680246490927, 'recall': 0.9184505606523955, 'f1': 0.9218963165075034, 'number': 2943}	{'precision': 0.9455665867853474, 'recall': 0.938817131203263, 'f1': 0.9421797714480641, 'number': 2942}	0.9355	0.9286	0.9320	0.9981
0.0021	6.0	51042	0.0137	{'precision': 0.9104477611940298, 'recall': 0.9119945633707102, 'f1': 0.911220505856391, 'number': 2943}	{'precision': 0.9331523583305056, 'recall': 0.9347382732834806, 'f1': 0.9339446425539141, 'number': 2942}	0.9218	0.9234	0.9226	0.9978
0.0012	7.0	59549	0.0133	{'precision': 0.9154399178363574, 'recall': 0.90859667006456, 'f1': 0.912005457025921, 'number': 2943}	{'precision': 0.9397260273972603, 'recall': 0.9326988443235894, 'f1': 0.9361992494029341, 'number': 2942}	0.9276	0.9206	0.9241	0.9981
0.0013	8.0	68056	0.0194	{'precision': 0.9192886456908345, 'recall': 0.9133537206931702, 'f1': 0.9163115732060677, 'number': 2943}	{'precision': 0.9442353746151214, 'recall': 0.938137321549966, 'f1': 0.9411764705882352, 'number': 2942}	0.9318	0.9257	0.9287	0.9979
0.0007	9.0	76563	0.0143	{'precision': 0.9239945466939332, 'recall': 0.9211688752973156, 'f1': 0.9225795473881231, 'number': 2943}	{'precision': 0.9457892942379816, 'recall': 0.9428959891230455, 'f1': 0.9443404255319149, 'number': 2942}	0.9349	0.9320	0.9335	0.9982
0.0004	10.0	85070	0.0158	{'precision': 0.9251446070091868, 'recall': 0.9238871899422358, 'f1': 0.9245154709282557, 'number': 2943}	{'precision': 0.9455411844792376, 'recall': 0.9442556084296397, 'f1': 0.9448979591836736, 'number': 2942}	0.9353	0.9341	0.9347	0.9982

Framework versions

• Transformers 4.26.0
• Pytorch 1.12.1
• Datasets 2.9.0
• Tokenizers 0.13.2