EElayoutlmv3_jordyvl_rvl_cdip_100_examples_per_class_2023-12-05_txt_vis_concat_enc_10_gate

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: 1.0000
• Accuracy: 0.75
• Exit 0 Accuracy: 0.055
• Exit 1 Accuracy: 0.22

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: 24
• total_train_batch_size: 192
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 60

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy	Exit 0 Accuracy	Exit 1 Accuracy
No log	0.96	4	2.7552	0.09	0.0425	0.0625
No log	1.96	8	2.7092	0.15	0.0475	0.0625
No log	2.96	12	2.6218	0.1825	0.0525	0.0625
No log	3.96	16	2.5483	0.1925	0.0525	0.0625
No log	4.96	20	2.4980	0.21	0.0525	0.0625
No log	5.96	24	2.3901	0.28	0.0525	0.0625
No log	6.96	28	2.2958	0.33	0.055	0.0625
No log	7.96	32	2.2144	0.34	0.055	0.0625
No log	8.96	36	2.1107	0.37	0.055	0.0625
No log	9.96	40	1.9969	0.405	0.055	0.0625
No log	10.96	44	1.8919	0.46	0.055	0.0625
No log	11.96	48	1.7897	0.4975	0.055	0.0625
No log	12.96	52	1.6686	0.525	0.055	0.0625
No log	13.96	56	1.6167	0.555	0.055	0.0625
No log	14.96	60	1.4750	0.605	0.055	0.0625
No log	15.96	64	1.4324	0.6225	0.055	0.065
No log	16.96	68	1.3211	0.645	0.055	0.0925
No log	17.96	72	1.2686	0.6675	0.055	0.1025
No log	18.96	76	1.2206	0.6725	0.055	0.115
No log	19.96	80	1.1536	0.7025	0.055	0.115
No log	20.96	84	1.1113	0.71	0.0525	0.115
No log	21.96	88	1.0655	0.715	0.0525	0.1175
No log	22.96	92	1.0423	0.735	0.0525	0.12
No log	23.96	96	1.0043	0.735	0.0525	0.1175
No log	24.96	100	1.0017	0.74	0.0525	0.12
No log	25.96	104	1.0167	0.7175	0.0525	0.12
No log	26.96	108	0.9570	0.74	0.0525	0.1175
No log	27.96	112	0.9620	0.7425	0.0525	0.12
No log	28.96	116	0.9466	0.7425	0.0525	0.1175
No log	29.96	120	0.9441	0.7575	0.0525	0.12
No log	30.96	124	0.9568	0.7375	0.0525	0.1175
No log	31.96	128	0.9313	0.7525	0.0525	0.11
No log	32.96	132	0.9330	0.74	0.0525	0.1025
No log	33.96	136	0.9370	0.76	0.0525	0.12
No log	34.96	140	0.9455	0.76	0.0525	0.1125
No log	35.96	144	0.9459	0.7625	0.0525	0.1025
No log	36.96	148	0.9418	0.7575	0.0525	0.0975
No log	37.96	152	0.9352	0.755	0.0525	0.105
No log	38.96	156	0.9377	0.7425	0.0525	0.1125
No log	39.96	160	0.9341	0.7525	0.0525	0.1175
No log	40.96	164	0.9452	0.7575	0.055	0.1475
No log	41.96	168	0.9486	0.7575	0.055	0.175
No log	42.96	172	0.9656	0.7525	0.055	0.1375
No log	43.96	176	0.9723	0.7525	0.0575	0.1575
No log	44.96	180	0.9682	0.75	0.0575	0.1775
No log	45.96	184	0.9699	0.7575	0.0575	0.195
No log	46.96	188	0.9695	0.7575	0.0575	0.1925
No log	47.96	192	0.9850	0.75	0.0575	0.1975
No log	48.96	196	0.9909	0.7575	0.0575	0.2075
No log	49.96	200	0.9751	0.75	0.0575	0.205
No log	50.96	204	0.9723	0.7525	0.0575	0.205
No log	51.96	208	0.9829	0.75	0.0575	0.21
No log	52.96	212	0.9833	0.755	0.0575	0.21
No log	53.96	216	0.9789	0.7575	0.0575	0.2125
No log	54.96	220	0.9781	0.7575	0.0575	0.2175
No log	55.96	224	0.9853	0.755	0.0575	0.2225
No log	56.96	228	0.9910	0.7525	0.0575	0.225
No log	57.96	232	0.9973	0.75	0.055	0.2225
No log	58.96	236	1.0001	0.75	0.055	0.2225
No log	59.96	240	1.0000	0.75	0.055	0.22

Framework versions

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