layoutlmv2_docvqa_first

This model is a fine-tuned version of microsoft/layoutlmv2-base-uncased on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 3.2572

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: 5e-05
• train_batch_size: 2
• eval_batch_size: 8
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 1

Training results

Training Loss	Epoch	Step	Validation Loss
3.8046	0.0221	10	4.6015
3.8232	0.0442	20	4.7682
3.7611	0.0664	30	4.6595
4.1572	0.0885	40	4.3268
4.1477	0.1106	50	4.0399
4.6312	0.1327	60	4.4170
4.2506	0.1549	70	3.8923
3.814	0.1770	80	4.2276
4.517	0.1991	90	4.2349
4.5446	0.2212	100	3.7886
4.482	0.2434	110	3.7871
4.0186	0.2655	120	3.8549
3.6858	0.2876	130	3.8050
3.826	0.3097	140	3.7714
3.6791	0.3319	150	4.1984
4.1924	0.3540	160	3.7813
3.8497	0.3761	170	3.6755
3.6868	0.3982	180	3.8626
4.4311	0.4204	190	3.6253
3.7333	0.4425	200	3.5505
3.971	0.4646	210	3.6746
3.5623	0.4867	220	3.4512
3.3989	0.5088	230	3.3751
3.5899	0.5310	240	3.3715
2.9724	0.5531	250	3.5085
4.0829	0.5752	260	3.4023
3.4239	0.5973	270	3.2687
3.3833	0.6195	280	3.4527
4.1309	0.6416	290	3.5902
3.7569	0.6637	300	3.3229
3.5462	0.6858	310	3.3342
3.4489	0.7080	320	3.3608
2.9949	0.7301	330	3.3340
3.3155	0.7522	340	3.3299
3.2365	0.7743	350	3.4070
3.3671	0.7965	360	3.3916
3.3138	0.8186	370	3.2162
3.6234	0.8407	380	3.1907
3.7106	0.8628	390	3.2870
2.9935	0.8850	400	3.2960
3.4659	0.9071	410	3.2714
2.8749	0.9292	420	3.2891
3.378	0.9513	430	3.2819
3.1755	0.9735	440	3.2651
3.4684	0.9956	450	3.2572

Framework versions

• Transformers 4.44.2
• Pytorch 2.3.1+cu121
• Datasets 2.20.0
• Tokenizers 0.19.1