LayoutLMv3_maveriq_tobacco3482_2023-07-04_longer

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: 0.4933
• Accuracy: 0.915

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: 2e-05
• train_batch_size: 8
• eval_batch_size: 4
• seed: 42
• gradient_accumulation_steps: 32
• total_train_batch_size: 256
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 200

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
No log	0.96	3	2.1414	0.285
No log	1.96	6	2.0216	0.265
No log	2.96	9	1.9444	0.265
No log	3.96	12	1.8877	0.335
No log	4.96	15	1.8160	0.315
No log	5.96	18	1.7139	0.33
No log	6.96	21	1.6301	0.36
No log	7.96	24	1.5155	0.47
No log	8.96	27	1.4009	0.555
No log	9.96	30	1.3059	0.56
No log	10.96	33	1.1493	0.67
No log	11.96	36	1.0559	0.725
No log	12.96	39	0.9505	0.75
No log	13.96	42	0.8301	0.78
No log	14.96	45	0.7531	0.775
No log	15.96	48	0.7030	0.79
No log	16.96	51	0.6294	0.82
No log	17.96	54	0.5819	0.845
No log	18.96	57	0.5381	0.87
No log	19.96	60	0.4852	0.87
No log	20.96	63	0.4581	0.91
No log	21.96	66	0.4429	0.895
No log	22.96	69	0.4065	0.915
No log	23.96	72	0.4065	0.895
No log	24.96	75	0.3598	0.915
No log	25.96	78	0.3476	0.925
No log	26.96	81	0.3413	0.93
No log	27.96	84	0.3544	0.9
No log	28.96	87	0.3239	0.93
No log	29.96	90	0.3187	0.92
No log	30.96	93	0.3090	0.92
No log	31.96	96	0.3495	0.915
No log	32.96	99	0.3075	0.93
No log	33.96	102	0.3509	0.92
No log	34.96	105	0.3499	0.925
No log	35.96	108	0.3176	0.925
No log	36.96	111	0.3260	0.915
No log	37.96	114	0.3245	0.925
No log	38.96	117	0.3139	0.92
No log	39.96	120	0.3667	0.915
No log	40.96	123	0.3410	0.925
No log	41.96	126	0.3278	0.925
No log	42.96	129	0.3518	0.925
No log	43.96	132	0.3617	0.92
No log	44.96	135	0.3642	0.93
No log	45.96	138	0.3686	0.925
No log	46.96	141	0.3784	0.92
No log	47.96	144	0.3826	0.92
No log	48.96	147	0.3734	0.925
No log	49.96	150	0.3763	0.925
No log	50.96	153	0.3931	0.92
No log	51.96	156	0.3982	0.92
No log	52.96	159	0.3960	0.92
No log	53.96	162	0.3896	0.925
No log	54.96	165	0.3917	0.925
No log	55.96	168	0.4016	0.92
No log	56.96	171	0.4098	0.92
No log	57.96	174	0.4124	0.92
No log	58.96	177	0.4127	0.92
No log	59.96	180	0.4115	0.92
No log	60.96	183	0.4134	0.92
No log	61.96	186	0.4173	0.92
No log	62.96	189	0.4209	0.92
No log	63.96	192	0.4230	0.915
No log	64.96	195	0.4259	0.915
No log	65.96	198	0.4289	0.915
No log	66.96	201	0.4318	0.915
No log	67.96	204	0.4333	0.915
No log	68.96	207	0.4325	0.915
No log	69.96	210	0.4317	0.915
No log	70.96	213	0.4336	0.915
No log	71.96	216	0.4356	0.915
No log	72.96	219	0.4372	0.915
No log	73.96	222	0.4375	0.915
No log	74.96	225	0.4381	0.915
No log	75.96	228	0.4393	0.915
No log	76.96	231	0.4418	0.915
No log	77.96	234	0.4444	0.915
No log	78.96	237	0.4470	0.915
No log	79.96	240	0.4491	0.915
No log	80.96	243	0.4492	0.915
No log	81.96	246	0.4474	0.915
No log	82.96	249	0.4443	0.915
No log	83.96	252	0.4445	0.915
No log	84.96	255	0.4477	0.915
No log	85.96	258	0.4492	0.915
No log	86.96	261	0.4501	0.915
No log	87.96	264	0.4510	0.915
No log	88.96	267	0.4520	0.915
No log	89.96	270	0.4525	0.915
No log	90.96	273	0.4531	0.915
No log	91.96	276	0.4530	0.915
No log	92.96	279	0.4518	0.915
No log	93.96	282	0.4499	0.915
No log	94.96	285	0.4485	0.915
No log	95.96	288	0.4496	0.915
No log	96.96	291	0.4525	0.915
No log	97.96	294	0.4562	0.915
No log	98.96	297	0.4596	0.915
No log	99.96	300	0.4629	0.915
No log	100.96	303	0.4639	0.915
No log	101.96	306	0.4641	0.915
No log	102.96	309	0.4630	0.915
No log	103.96	312	0.4619	0.915
No log	104.96	315	0.4624	0.915
No log	105.96	318	0.4628	0.915
No log	106.96	321	0.4635	0.915
No log	107.96	324	0.4641	0.915
No log	108.96	327	0.4650	0.915
No log	109.96	330	0.4652	0.915
No log	110.96	333	0.4664	0.915
No log	111.96	336	0.4686	0.915
No log	112.96	339	0.4718	0.915
No log	113.96	342	0.4730	0.915
No log	114.96	345	0.4719	0.915
No log	115.96	348	0.4697	0.915
No log	116.96	351	0.4676	0.915
No log	117.96	354	0.4658	0.915
No log	118.96	357	0.4655	0.915
No log	119.96	360	0.4670	0.915
No log	120.96	363	0.4695	0.915
No log	121.96	366	0.4728	0.915
No log	122.96	369	0.4757	0.915
No log	123.96	372	0.4776	0.915
No log	124.96	375	0.4782	0.915
No log	125.96	378	0.4782	0.915
No log	126.96	381	0.4770	0.915
No log	127.96	384	0.4760	0.915
No log	128.96	387	0.4754	0.915
No log	129.96	390	0.4746	0.915
No log	130.96	393	0.4745	0.915
No log	131.96	396	0.4750	0.915
No log	132.96	399	0.4756	0.915
No log	133.96	402	0.4766	0.915
No log	134.96	405	0.4777	0.915
No log	135.96	408	0.4788	0.915
No log	136.96	411	0.4799	0.915
No log	137.96	414	0.4806	0.915
No log	138.96	417	0.4806	0.915
No log	139.96	420	0.4805	0.915
No log	140.96	423	0.4796	0.915
No log	141.96	426	0.4789	0.915
No log	142.96	429	0.4785	0.915
No log	143.96	432	0.4793	0.915
No log	144.96	435	0.4805	0.915
No log	145.96	438	0.4814	0.915
No log	146.96	441	0.4822	0.915
No log	147.96	444	0.4831	0.915
No log	148.96	447	0.4840	0.915
No log	149.96	450	0.4839	0.915
No log	150.96	453	0.4839	0.915
No log	151.96	456	0.4842	0.915
No log	152.96	459	0.4843	0.915
No log	153.96	462	0.4841	0.915
No log	154.96	465	0.4838	0.915
No log	155.96	468	0.4843	0.915
No log	156.96	471	0.4848	0.915
No log	157.96	474	0.4851	0.915
No log	158.96	477	0.4853	0.915
No log	159.96	480	0.4854	0.915
No log	160.96	483	0.4857	0.915
No log	161.96	486	0.4861	0.915
No log	162.96	489	0.4867	0.915
No log	163.96	492	0.4873	0.915
No log	164.96	495	0.4884	0.915
No log	165.96	498	0.4895	0.915
0.1894	166.96	501	0.4906	0.915
0.1894	167.96	504	0.4912	0.915
0.1894	168.96	507	0.4916	0.915
0.1894	169.96	510	0.4915	0.915
0.1894	170.96	513	0.4913	0.915
0.1894	171.96	516	0.4912	0.915
0.1894	172.96	519	0.4912	0.915
0.1894	173.96	522	0.4913	0.915
0.1894	174.96	525	0.4911	0.915
0.1894	175.96	528	0.4909	0.915
0.1894	176.96	531	0.4910	0.915
0.1894	177.96	534	0.4910	0.915
0.1894	178.96	537	0.4910	0.915
0.1894	179.96	540	0.4909	0.915
0.1894	180.96	543	0.4910	0.915
0.1894	181.96	546	0.4914	0.915
0.1894	182.96	549	0.4920	0.915
0.1894	183.96	552	0.4926	0.915
0.1894	184.96	555	0.4930	0.915
0.1894	185.96	558	0.4933	0.915
0.1894	186.96	561	0.4936	0.915
0.1894	187.96	564	0.4939	0.915
0.1894	188.96	567	0.4939	0.915
0.1894	189.96	570	0.4938	0.915
0.1894	190.96	573	0.4938	0.915
0.1894	191.96	576	0.4936	0.915
0.1894	192.96	579	0.4935	0.915
0.1894	193.96	582	0.4934	0.915
0.1894	194.96	585	0.4934	0.915
0.1894	195.96	588	0.4934	0.915
0.1894	196.96	591	0.4934	0.915
0.1894	197.96	594	0.4933	0.915
0.1894	198.96	597	0.4933	0.915
0.1894	199.96	600	0.4933	0.915

Framework versions

• Transformers 4.26.1
• Pytorch 1.13.1.post200
• Datasets 2.9.0
• Tokenizers 0.13.2