layoutlm-finetuned

This model is a fine-tuned version of impira/layoutlm-invoices on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 0.0064

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: 4
• eval_batch_size: 4
• seed: 42
• gradient_accumulation_steps: 4
• total_train_batch_size: 16
• optimizer: Use OptimizerNames.ADAMW_TORCH with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
• lr_scheduler_type: linear
• lr_scheduler_warmup_steps: 100
• num_epochs: 10
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss
0.0837	0.2116	100	0.0147
0.0603	0.4233	200	0.0101
0.0396	0.6349	300	0.0093
0.0442	0.8466	400	0.0086
0.0287	1.0571	500	0.0073
0.0396	1.2688	600	0.0084
0.0179	1.4804	700	0.0080
0.0391	1.6921	800	0.0069
0.0261	1.9037	900	0.0064
0.0146	2.1143	1000	0.0068
0.0192	2.3259	1100	0.0089
0.0233	2.5376	1200	0.0070
0.0129	2.7492	1300	0.0072
0.0173	2.9608	1400	0.0068
0.0217	3.1714	1500	0.0076
0.0198	3.3831	1600	0.0066
0.0163	3.5947	1700	0.0068
0.0142	3.8063	1800	0.0078
0.0144	4.0169	1900	0.0065
0.0143	4.2286	2000	0.0065
0.0125	4.4402	2100	0.0071
0.0108	4.6519	2200	0.0077
0.0138	4.8635	2300	0.0071
0.0146	5.0741	2400	0.0076
0.0135	5.2857	2500	0.0084
0.0213	5.4974	2600	0.0081
0.0117	5.7090	2700	0.0083
0.0181	5.9206	2800	0.0092
0.0124	6.1312	2900	0.0090
0.0102	6.3429	3000	0.0090
0.0110	6.5545	3100	0.0081
0.0100	6.7661	3200	0.0089
0.0111	6.9778	3300	0.0106
0.0127	7.1884	3400	0.0107
0.0075	7.4	3500	0.0106
0.0155	7.6116	3600	0.0104
0.0114	7.8233	3700	0.0104
0.0084	8.0339	3800	0.0100
0.0075	8.2455	3900	0.0112
0.0095	8.4571	4000	0.0115
0.0117	8.6688	4100	0.0115
0.0083	8.8804	4200	0.0118
0.0060	9.0910	4300	0.0118
0.0049	9.3026	4400	0.0130
0.0082	9.5143	4500	0.0129
0.0083	9.7259	4600	0.0131
0.0076	9.9376	4700	0.0130
0.0062	10.0	4730	0.0130

Framework versions

• Transformers 5.11.0
• Pytorch 2.6.0+cu124
• Datasets 5.0.0
• Tokenizers 0.22.2