layoutlmv3-finetuned-wildreceipt

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: 0.3146
• Precision: 0.8798
• Recall: 0.8807
• F1: 0.8802
• Accuracy: 0.9260

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: 4
• eval_batch_size: 4
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• training_steps: 4000

Training results

Training Loss	Epoch	Step	Validation Loss	Precision	Recall	F1	Accuracy
No log	0.3155	100	1.3209	0.5494	0.3560	0.4320	0.6831
No log	0.6309	200	0.8725	0.6581	0.5196	0.5807	0.7703
No log	0.9464	300	0.6947	0.7348	0.6701	0.7009	0.8309
No log	1.2618	400	0.5767	0.7727	0.7133	0.7418	0.8519
1.0485	1.5773	500	0.5011	0.7667	0.7605	0.7636	0.8673
1.0485	1.8927	600	0.4996	0.7616	0.7818	0.7716	0.8668
1.0485	2.2082	700	0.4144	0.8010	0.7989	0.7999	0.8867
1.0485	2.5237	800	0.3943	0.8162	0.8231	0.8196	0.8936
1.0485	2.8391	900	0.3821	0.8096	0.8340	0.8216	0.8963
0.418	3.1546	1000	0.3637	0.8350	0.8428	0.8389	0.9031
0.418	3.4700	1100	0.3422	0.8518	0.8314	0.8415	0.9081
0.418	3.7855	1200	0.3425	0.8322	0.8579	0.8449	0.9057
0.418	4.1009	1300	0.3242	0.8556	0.8541	0.8549	0.9132
0.418	4.4164	1400	0.3267	0.8442	0.8628	0.8534	0.9117
0.28	4.7319	1500	0.3144	0.8565	0.8603	0.8584	0.9152
0.28	5.0473	1600	0.3102	0.8707	0.8617	0.8661	0.9182
0.28	5.3628	1700	0.3291	0.8677	0.8615	0.8646	0.9169
0.28	5.6782	1800	0.2999	0.8587	0.8717	0.8652	0.9182
0.28	5.9937	1900	0.3128	0.8596	0.8680	0.8638	0.9184
0.2106	6.3091	2000	0.3111	0.8667	0.8725	0.8696	0.9207
0.2106	6.6246	2100	0.3053	0.8716	0.8679	0.8697	0.9207
0.2106	6.9401	2200	0.3017	0.8683	0.8756	0.8719	0.9221
0.2106	7.2555	2300	0.2986	0.8800	0.8702	0.8751	0.9236
0.2106	7.5710	2400	0.3012	0.8747	0.8717	0.8732	0.9230
0.1724	7.8864	2500	0.3014	0.8696	0.8776	0.8736	0.9232
0.1724	8.2019	2600	0.3066	0.8763	0.8725	0.8744	0.9232
0.1724	8.5174	2700	0.3181	0.8696	0.8794	0.8745	0.9228
0.1724	8.8328	2800	0.3065	0.8724	0.8816	0.8770	0.9237
0.1724	9.1483	2900	0.3041	0.8753	0.8818	0.8785	0.9259
0.1393	9.4637	3000	0.3205	0.8685	0.8812	0.8748	0.9238
0.1393	9.7792	3100	0.3050	0.8705	0.8838	0.8771	0.9253
0.1393	10.0946	3200	0.3158	0.8746	0.8838	0.8792	0.9246
0.1393	10.4101	3300	0.3086	0.8785	0.8779	0.8782	0.9258
0.1393	10.7256	3400	0.3108	0.8801	0.8822	0.8811	0.9274
0.1222	11.0410	3500	0.3146	0.8798	0.8807	0.8802	0.9260
0.1222	11.3565	3600	0.3111	0.8790	0.8813	0.8802	0.9261
0.1222	11.6719	3700	0.3161	0.8769	0.8848	0.8808	0.9261
0.1222	11.9874	3800	0.3133	0.8769	0.8823	0.8796	0.9262
0.1222	12.3028	3900	0.3137	0.8767	0.8829	0.8798	0.9260
0.1115	12.6183	4000	0.3135	0.8766	0.8833	0.8799	0.9262

Framework versions

• Transformers 4.41.0.dev0
• Pytorch 2.2.1+cu121
• Datasets 2.19.0
• Tokenizers 0.19.1