layoutlm-captive-corp-160-init

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

• Loss: 0.3305
• Precision: 0.8602
• Recall: 0.8843
• F1: 0.8721
• Accuracy: 0.9428

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: 1e-05
• train_batch_size: 2
• eval_batch_size: 2
• seed: 42
• 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
• num_epochs: 50
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Precision	Recall	F1	Accuracy
3.6934	1.0	33	2.9425	0.0	0.0	0.0	0.4747
2.8238	2.0	66	2.3730	0.2758	0.3864	0.3219	0.5851
2.3806	3.0	99	2.0655	0.2810	0.3940	0.3280	0.5948
2.1004	4.0	132	1.8309	0.4008	0.4080	0.4044	0.6436
1.8787	5.0	165	1.6368	0.5	0.4795	0.4896	0.6996
1.7022	6.0	198	1.4801	0.5685	0.5587	0.5635	0.7468
1.5448	7.0	231	1.3572	0.6253	0.6340	0.6296	0.7878
1.4106	8.0	264	1.2286	0.6565	0.6728	0.6645	0.8102
1.2955	9.0	297	1.1187	0.6891	0.7110	0.6999	0.8363
1.1797	10.0	330	1.0272	0.7074	0.7298	0.7184	0.8462
1.0856	11.0	363	0.9504	0.7393	0.7648	0.7519	0.8651
1.002	12.0	396	0.8621	0.7794	0.7987	0.7889	0.8878
0.922	13.0	429	0.8164	0.7941	0.8181	0.8059	0.8950
0.8502	14.0	462	0.7482	0.8052	0.8278	0.8163	0.9078
0.7872	15.0	495	0.7002	0.8090	0.8321	0.8204	0.9141
0.725	16.0	528	0.6489	0.8229	0.8477	0.8351	0.9225
0.6719	17.0	561	0.6104	0.8291	0.8515	0.8401	0.9277
0.6278	18.0	594	0.5823	0.8352	0.8590	0.8469	0.9287
0.5794	19.0	627	0.5377	0.8383	0.8595	0.8488	0.9341
0.542	20.0	660	0.5138	0.8445	0.8622	0.8533	0.9365
0.5025	21.0	693	0.5009	0.8408	0.8644	0.8524	0.9333
0.4722	22.0	726	0.4753	0.8424	0.8687	0.8553	0.9367
0.4489	23.0	759	0.4546	0.8429	0.8665	0.8546	0.9373
0.4195	24.0	792	0.4425	0.8462	0.8703	0.8581	0.9375
0.4027	25.0	825	0.4263	0.8514	0.8730	0.8621	0.9394
0.3803	26.0	858	0.4214	0.8456	0.8698	0.8575	0.9355
0.3666	27.0	891	0.4129	0.8480	0.8735	0.8606	0.9375
0.3467	28.0	924	0.3990	0.8509	0.8724	0.8615	0.9380
0.3308	29.0	957	0.3934	0.8567	0.8751	0.8658	0.9388
0.3188	30.0	990	0.3916	0.8446	0.8746	0.8593	0.9361
0.3102	31.0	1023	0.3761	0.8526	0.8746	0.8634	0.9382
0.3002	32.0	1056	0.3795	0.8539	0.8773	0.8654	0.9392
0.2868	33.0	1089	0.3643	0.8520	0.8741	0.8629	0.9394
0.2759	34.0	1122	0.3473	0.8550	0.8789	0.8668	0.9440
0.2685	35.0	1155	0.3520	0.8542	0.8800	0.8669	0.9408
0.2637	36.0	1188	0.3489	0.8549	0.8784	0.8665	0.9394
0.2553	37.0	1221	0.3515	0.8503	0.8805	0.8652	0.9398
0.2503	38.0	1254	0.3464	0.8536	0.8816	0.8674	0.9408
0.2462	39.0	1287	0.3457	0.8548	0.8805	0.8674	0.9404
0.2413	40.0	1320	0.3417	0.8552	0.8805	0.8677	0.9408
0.2328	41.0	1353	0.3338	0.8580	0.8816	0.8697	0.9426
0.2341	42.0	1386	0.3336	0.8563	0.8821	0.8690	0.9428
0.2295	43.0	1419	0.3405	0.8521	0.8805	0.8661	0.9404
0.2232	44.0	1452	0.3369	0.8561	0.8805	0.8681	0.9410
0.2244	45.0	1485	0.3314	0.8562	0.8811	0.8684	0.9416
0.2185	46.0	1518	0.3340	0.8571	0.8816	0.8692	0.9420
0.2175	47.0	1551	0.3317	0.8595	0.8827	0.8710	0.9422
0.2142	48.0	1584	0.3305	0.8602	0.8843	0.8721	0.9428
0.2162	49.0	1617	0.3333	0.8558	0.8816	0.8685	0.9410
0.2177	50.0	1650	0.3323	0.8586	0.8827	0.8705	0.9420

Framework versions

• Transformers 4.49.0.dev0
• Pytorch 2.5.1+cu121
• Datasets 3.2.0
• Tokenizers 0.21.0