EElayoutlmv3_jordyvl_rvl_cdip_100_examples_per_class_2023-09-10_txt_vis_concat_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: 0.9407
• Accuracy: 0.78
• Exit 0 Accuracy: 0.0625

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
No log	0.96	16	2.6995	0.13	0.0625
No log	1.98	33	2.5653	0.21	0.06
No log	3.0	50	2.3927	0.2825	0.07
No log	3.96	66	2.2103	0.345	0.075
No log	4.98	83	2.0217	0.4525	0.065
No log	6.0	100	1.8175	0.5325	0.06
No log	6.96	116	1.6096	0.5875	0.0625
No log	7.98	133	1.4160	0.6375	0.0625
No log	9.0	150	1.3283	0.6575	0.0625
No log	9.96	166	1.2253	0.7	0.0625
No log	10.98	183	1.1531	0.7225	0.0625
No log	12.0	200	1.0661	0.7375	0.0625
No log	12.96	216	1.0565	0.73	0.0625
No log	13.98	233	1.0281	0.73	0.0625
No log	15.0	250	1.0459	0.7275	0.0625
No log	15.96	266	0.9802	0.75	0.0625
No log	16.98	283	0.9665	0.7525	0.0625
No log	18.0	300	0.9655	0.7475	0.0625
No log	18.96	316	0.9463	0.7675	0.0625
No log	19.98	333	0.9392	0.765	0.0625
No log	21.0	350	0.9768	0.75	0.0625
No log	21.96	366	0.9973	0.7525	0.0625
No log	22.98	383	0.9660	0.765	0.0625
No log	24.0	400	1.0065	0.7475	0.0625
No log	24.96	416	0.9077	0.7825	0.0625
No log	25.98	433	0.9568	0.775	0.0625
No log	27.0	450	0.9389	0.775	0.0625
No log	27.96	466	0.9266	0.78	0.0625
No log	28.98	483	0.9301	0.7825	0.0625
0.5845	30.0	500	0.9220	0.785	0.0625
0.5845	30.96	516	0.9563	0.77	0.0625
0.5845	31.98	533	0.9272	0.785	0.0625
0.5845	33.0	550	0.9430	0.7775	0.0625
0.5845	33.96	566	0.9525	0.78	0.0625
0.5845	34.98	583	0.9190	0.7975	0.0625
0.5845	36.0	600	0.9416	0.765	0.0625
0.5845	36.96	616	0.9286	0.7825	0.0625
0.5845	37.98	633	0.9411	0.775	0.0625
0.5845	39.0	650	0.9468	0.77	0.0625
0.5845	39.96	666	0.9305	0.7825	0.0625
0.5845	40.98	683	0.9428	0.775	0.0625
0.5845	42.0	700	0.9484	0.78	0.0625
0.5845	42.96	716	0.9411	0.7825	0.0625
0.5845	43.98	733	0.9564	0.775	0.0625
0.5845	45.0	750	0.9293	0.785	0.0625
0.5845	45.96	766	0.9578	0.78	0.0625
0.5845	46.98	783	0.9377	0.79	0.0625
0.5845	48.0	800	0.9417	0.78	0.0625
0.5845	48.96	816	0.9495	0.7825	0.0625
0.5845	49.98	833	0.9401	0.7875	0.0625
0.5845	51.0	850	0.9458	0.7875	0.0625
0.5845	51.96	866	0.9468	0.7875	0.0625
0.5845	52.98	883	0.9341	0.79	0.0625
0.5845	54.0	900	0.9344	0.7875	0.0625
0.5845	54.96	916	0.9350	0.785	0.0625
0.5845	55.98	933	0.9391	0.78	0.0625
0.5845	57.0	950	0.9408	0.78	0.0625
0.5845	57.6	960	0.9407	0.78	0.0625

Framework versions

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