layoutlm-funsd

This model is a fine-tuned version of microsoft/layoutlm-base-uncased on the funsd dataset. It achieves the following results on the evaluation set:

• Loss: 1.1243
• Answer: {'precision': 0.40076335877862596, 'recall': 0.519159456118665, 'f1': 0.4523424878836834, 'number': 809}
• Header: {'precision': 0.28421052631578947, 'recall': 0.226890756302521, 'f1': 0.25233644859813087, 'number': 119}
• Question: {'precision': 0.5280065897858319, 'recall': 0.6018779342723005, 'f1': 0.5625274243089073, 'number': 1065}
• Overall Precision: 0.4616
• Overall Recall: 0.5459
• Overall F1: 0.5002
• Overall Accuracy: 0.6215

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: 16
• eval_batch_size: 8
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 15
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Answer	Header	Question	Overall Precision	Overall Recall	Overall F1	Overall Accuracy
1.7728	1.0	10	1.5441	{'precision': 0.04580152671755725, 'recall': 0.059332509270704575, 'f1': 0.05169628432956382, 'number': 809}	{'precision': 0.0, 'recall': 0.0, 'f1': 0.0, 'number': 119}	{'precision': 0.20335429769392033, 'recall': 0.18215962441314554, 'f1': 0.19217434373452202, 'number': 1065}	0.1209	0.1214	0.1212	0.3719
1.4551	2.0	20	1.3517	{'precision': 0.20478234212139793, 'recall': 0.41285537700865266, 'f1': 0.27377049180327867, 'number': 809}	{'precision': 0.0, 'recall': 0.0, 'f1': 0.0, 'number': 119}	{'precision': 0.26090225563909775, 'recall': 0.32582159624413143, 'f1': 0.28977035490605424, 'number': 1065}	0.2297	0.3417	0.2747	0.4263
1.295	3.0	30	1.2465	{'precision': 0.26224426534407935, 'recall': 0.522867737948084, 'f1': 0.34929810074318746, 'number': 809}	{'precision': 0.058823529411764705, 'recall': 0.01680672268907563, 'f1': 0.026143790849673203, 'number': 119}	{'precision': 0.3458528951486698, 'recall': 0.41502347417840374, 'f1': 0.37729406743491256, 'number': 1065}	0.2964	0.4350	0.3526	0.4803
1.1635	4.0	40	1.1449	{'precision': 0.28778467908902694, 'recall': 0.515451174289246, 'f1': 0.3693534100974314, 'number': 809}	{'precision': 0.2638888888888889, 'recall': 0.15966386554621848, 'f1': 0.19895287958115182, 'number': 119}	{'precision': 0.412396694214876, 'recall': 0.46854460093896716, 'f1': 0.4386813186813187, 'number': 1065}	0.3424	0.4691	0.3959	0.5521
1.0456	5.0	50	1.0703	{'precision': 0.3060240963855422, 'recall': 0.47095179233621753, 'f1': 0.37098344693281404, 'number': 809}	{'precision': 0.3472222222222222, 'recall': 0.21008403361344538, 'f1': 0.2617801047120419, 'number': 119}	{'precision': 0.40298507462686567, 'recall': 0.5830985915492958, 'f1': 0.476592478894858, 'number': 1065}	0.3593	0.5153	0.4234	0.5797
0.9601	6.0	60	1.2304	{'precision': 0.30907920154539603, 'recall': 0.5933250927070457, 'f1': 0.40643522438611346, 'number': 809}	{'precision': 0.3333333333333333, 'recall': 0.16806722689075632, 'f1': 0.223463687150838, 'number': 119}	{'precision': 0.4642857142857143, 'recall': 0.4394366197183099, 'f1': 0.4515195369030391, 'number': 1065}	0.3693	0.4857	0.4196	0.5479
0.9153	7.0	70	1.1091	{'precision': 0.35518157661647476, 'recall': 0.4956736711990111, 'f1': 0.41382868937048506, 'number': 809}	{'precision': 0.3125, 'recall': 0.21008403361344538, 'f1': 0.25125628140703515, 'number': 119}	{'precision': 0.5262645914396887, 'recall': 0.507981220657277, 'f1': 0.5169612995699953, 'number': 1065}	0.4323	0.4852	0.4572	0.6011
0.8346	8.0	80	1.0632	{'precision': 0.35597826086956524, 'recall': 0.4857849196538937, 'f1': 0.4108729743857816, 'number': 809}	{'precision': 0.28421052631578947, 'recall': 0.226890756302521, 'f1': 0.25233644859813087, 'number': 119}	{'precision': 0.46401799100449775, 'recall': 0.5812206572769953, 'f1': 0.516048353480617, 'number': 1065}	0.4102	0.5213	0.4591	0.6103
0.7789	9.0	90	1.0955	{'precision': 0.3817062445030783, 'recall': 0.5364647713226205, 'f1': 0.44604316546762585, 'number': 809}	{'precision': 0.26, 'recall': 0.2184873949579832, 'f1': 0.23744292237442924, 'number': 119}	{'precision': 0.5137693631669535, 'recall': 0.5605633802816902, 'f1': 0.5361472833408173, 'number': 1065}	0.4406	0.5304	0.4813	0.6082
0.7751	10.0	100	1.1232	{'precision': 0.38474434199497065, 'recall': 0.5673671199011124, 'f1': 0.45854145854145856, 'number': 809}	{'precision': 0.3010752688172043, 'recall': 0.23529411764705882, 'f1': 0.2641509433962264, 'number': 119}	{'precision': 0.5040358744394619, 'recall': 0.5276995305164319, 'f1': 0.5155963302752293, 'number': 1065}	0.4369	0.5263	0.4775	0.6032
0.6875	11.0	110	1.1092	{'precision': 0.39342723004694835, 'recall': 0.5179233621755254, 'f1': 0.44717182497331914, 'number': 809}	{'precision': 0.34146341463414637, 'recall': 0.23529411764705882, 'f1': 0.27860696517412936, 'number': 119}	{'precision': 0.5076305220883535, 'recall': 0.5934272300469483, 'f1': 0.5471861471861472, 'number': 1065}	0.4511	0.5414	0.4921	0.6233
0.6808	12.0	120	1.1286	{'precision': 0.40641158221303, 'recall': 0.4857849196538937, 'f1': 0.44256756756756754, 'number': 809}	{'precision': 0.24561403508771928, 'recall': 0.23529411764705882, 'f1': 0.24034334763948498, 'number': 119}	{'precision': 0.49772036474164133, 'recall': 0.6150234741784038, 'f1': 0.5501889962200757, 'number': 1065}	0.4489	0.5399	0.4902	0.6159
0.656	13.0	130	1.1237	{'precision': 0.39822134387351776, 'recall': 0.49814585908529047, 'f1': 0.442613948380011, 'number': 809}	{'precision': 0.2967032967032967, 'recall': 0.226890756302521, 'f1': 0.2571428571428572, 'number': 119}	{'precision': 0.5141732283464567, 'recall': 0.6131455399061033, 'f1': 0.5593147751605996, 'number': 1065}	0.4564	0.5434	0.4961	0.6179
0.6359	14.0	140	1.1296	{'precision': 0.3996399639963996, 'recall': 0.5488257107540173, 'f1': 0.46249999999999997, 'number': 809}	{'precision': 0.32926829268292684, 'recall': 0.226890756302521, 'f1': 0.26865671641791045, 'number': 119}	{'precision': 0.5376712328767124, 'recall': 0.5896713615023474, 'f1': 0.5624720107478729, 'number': 1065}	0.4655	0.5514	0.5048	0.6173
0.6117	15.0	150	1.1243	{'precision': 0.40076335877862596, 'recall': 0.519159456118665, 'f1': 0.4523424878836834, 'number': 809}	{'precision': 0.28421052631578947, 'recall': 0.226890756302521, 'f1': 0.25233644859813087, 'number': 119}	{'precision': 0.5280065897858319, 'recall': 0.6018779342723005, 'f1': 0.5625274243089073, 'number': 1065}	0.4616	0.5459	0.5002	0.6215

Framework versions

• Transformers 4.38.2
• Pytorch 2.1.0+cu121
• Datasets 2.18.0
• Tokenizers 0.15.2