EElayoutlmv3_jordyvl_rvl_cdip_100_examples_per_class_2023-08-21_text_vision_enc_1_2_3_4_ramp

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: 1.5044
• Accuracy: 0.74
• Exit 0 Accuracy: 0.075
• Exit 1 Accuracy: 0.0925
• Exit 2 Accuracy: 0.44
• Exit 3 Accuracy: 0.525
• Exit 4 Accuracy: 0.5675
• Exit 5 Accuracy: 0.635

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: 2
• eval_batch_size: 1
• seed: 42
• gradient_accumulation_steps: 24
• total_train_batch_size: 48
• 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	Exit 2 Accuracy	Exit 3 Accuracy	Exit 4 Accuracy	Exit 5 Accuracy
No log	0.96	16	2.6509	0.18	0.075	0.0925	0.0625	0.0625	0.0625	0.0625
No log	1.98	33	2.4283	0.2925	0.0775	0.0925	0.0625	0.0625	0.0625	0.0625
No log	3.0	50	2.2343	0.3325	0.0875	0.0925	0.0625	0.0625	0.0625	0.0625
No log	3.96	66	1.9465	0.4425	0.08	0.0925	0.0625	0.0625	0.0625	0.0625
No log	4.98	83	1.6496	0.6025	0.075	0.0925	0.0625	0.0625	0.0625	0.0625
No log	6.0	100	1.4404	0.635	0.075	0.0925	0.0625	0.0625	0.0625	0.0625
No log	6.96	116	1.2292	0.715	0.08	0.0925	0.065	0.0625	0.0625	0.0625
No log	7.98	133	1.0856	0.715	0.075	0.0925	0.065	0.0625	0.0625	0.0625
No log	9.0	150	0.9673	0.7475	0.0775	0.0925	0.065	0.0625	0.0625	0.0625
No log	9.96	166	0.9172	0.7475	0.075	0.0925	0.065	0.0625	0.0625	0.0625
No log	10.98	183	0.9036	0.7525	0.0775	0.0925	0.065	0.0625	0.0625	0.0625
No log	12.0	200	0.8164	0.7775	0.075	0.0925	0.065	0.0625	0.0625	0.0625
No log	12.96	216	0.8499	0.765	0.0775	0.0925	0.07	0.0625	0.0625	0.0625
No log	13.98	233	0.9053	0.775	0.0775	0.0925	0.07	0.0625	0.0625	0.0625
No log	15.0	250	0.9470	0.775	0.08	0.0925	0.0675	0.0625	0.0625	0.0625
No log	15.96	266	0.9509	0.7575	0.075	0.0925	0.0675	0.0625	0.0625	0.0625
No log	16.98	283	0.9221	0.78	0.075	0.0925	0.0725	0.0625	0.0625	0.07
No log	18.0	300	0.9725	0.775	0.0775	0.0925	0.0925	0.0625	0.0625	0.1025
No log	18.96	316	1.1409	0.7625	0.0825	0.0925	0.125	0.0625	0.0625	0.11
No log	19.98	333	1.0653	0.7825	0.0925	0.0925	0.1475	0.0625	0.0625	0.1425
No log	21.0	350	1.0736	0.78	0.0775	0.0925	0.1875	0.0625	0.0625	0.15
No log	21.96	366	1.1706	0.7725	0.075	0.0925	0.2275	0.065	0.0575	0.1975
No log	22.98	383	1.1940	0.76	0.0775	0.0925	0.215	0.1325	0.145	0.2
No log	24.0	400	1.0195	0.7875	0.0775	0.0925	0.2225	0.19	0.17	0.2575
No log	24.96	416	1.1589	0.7625	0.075	0.0925	0.2725	0.26	0.2425	0.29
No log	25.98	433	1.2225	0.75	0.075	0.0925	0.255	0.2825	0.2875	0.3375
No log	27.0	450	1.1789	0.7575	0.0775	0.0925	0.2775	0.32	0.3775	0.3775
No log	27.96	466	1.1574	0.7725	0.075	0.09	0.29	0.3375	0.3775	0.4225
No log	28.98	483	1.2567	0.7525	0.09	0.0925	0.3375	0.3575	0.43	0.42
1.6703	30.0	500	1.1840	0.7775	0.09	0.0925	0.35	0.375	0.44	0.45
1.6703	30.96	516	1.2607	0.7575	0.08	0.0925	0.3375	0.375	0.4475	0.4675
1.6703	31.98	533	1.2006	0.775	0.0775	0.0925	0.345	0.395	0.4525	0.47
1.6703	33.0	550	1.3099	0.745	0.0775	0.0925	0.32	0.38	0.4575	0.4725
1.6703	33.96	566	1.2074	0.77	0.075	0.0925	0.37	0.4275	0.4825	0.5175
1.6703	34.98	583	1.2929	0.76	0.075	0.0925	0.375	0.4375	0.48	0.5125
1.6703	36.0	600	1.2752	0.7625	0.0775	0.0925	0.395	0.4575	0.52	0.5125
1.6703	36.96	616	1.3596	0.7475	0.0825	0.0925	0.4	0.47	0.5075	0.52
1.6703	37.98	633	1.3920	0.735	0.0775	0.0925	0.3925	0.4725	0.52	0.5275
1.6703	39.0	650	1.4005	0.7475	0.075	0.0925	0.3875	0.455	0.5125	0.5225
1.6703	39.96	666	1.3938	0.75	0.0775	0.0925	0.415	0.4925	0.555	0.54
1.6703	40.98	683	1.3711	0.755	0.0775	0.0925	0.4075	0.4775	0.5375	0.57
1.6703	42.0	700	1.3591	0.7475	0.075	0.0925	0.415	0.51	0.5525	0.575
1.6703	42.96	716	1.4183	0.7475	0.0775	0.0925	0.4275	0.5	0.555	0.5875
1.6703	43.98	733	1.3572	0.7475	0.075	0.0925	0.4275	0.505	0.55	0.595
1.6703	45.0	750	1.4095	0.755	0.0775	0.0925	0.4225	0.5075	0.56	0.6025
1.6703	45.96	766	1.4217	0.7425	0.0775	0.0925	0.435	0.5	0.56	0.5975
1.6703	46.98	783	1.3684	0.76	0.075	0.0925	0.44	0.5075	0.56	0.6075
1.6703	48.0	800	1.4222	0.7625	0.075	0.0925	0.4425	0.5125	0.57	0.62
1.6703	48.96	816	1.4734	0.75	0.075	0.0925	0.435	0.5175	0.57	0.615
1.6703	49.98	833	1.4910	0.7425	0.075	0.0925	0.44	0.515	0.5625	0.62
1.6703	51.0	850	1.4879	0.7525	0.075	0.0925	0.445	0.52	0.57	0.6125
1.6703	51.96	866	1.4889	0.7525	0.075	0.0925	0.445	0.5275	0.5775	0.615
1.6703	52.98	883	1.4836	0.75	0.075	0.0925	0.4425	0.5175	0.57	0.6275
1.6703	54.0	900	1.4904	0.7475	0.075	0.0925	0.44	0.52	0.5725	0.6275
1.6703	54.96	916	1.4920	0.745	0.075	0.0925	0.4425	0.5225	0.5725	0.63
1.6703	55.98	933	1.5048	0.7425	0.075	0.0925	0.44	0.5225	0.5675	0.6325
1.6703	57.0	950	1.5058	0.74	0.075	0.0925	0.44	0.525	0.5675	0.635
1.6703	57.6	960	1.5044	0.74	0.075	0.0925	0.44	0.525	0.5675	0.635

Framework versions

• Transformers 4.31.0
• Pytorch 2.0.1+cu117
• Datasets 2.13.1
• Tokenizers 0.13.3