Layout-finetuned-fr-model-107instances107-150epochs-5e-05lr-GPU

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

• Loss: 0.0000
• Accuracy: 1.0
• Learning Rate: 1e-05

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: 4
• eval_batch_size: 8
• seed: 42
• optimizer: Use OptimizerNames.ADAMW_TORCH with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
• lr_scheduler_type: reduce_lr_on_plateau
• lr_scheduler_warmup_ratio: 0.06
• num_epochs: 150

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy	Learning Rate
0.0236	3.7037	100	0.0071	0.9953	5e-05
0.0547	7.4074	200	0.1518	0.9813	5e-05
0.1261	11.1111	300	0.0840	0.9860	5e-05
0.1562	14.8148	400	0.1556	0.9860	5e-05
0.0973	18.5185	500	0.0426	0.9907	5e-05
0.0273	22.2222	600	0.0005	1.0	5e-05
0.0709	25.9259	700	0.0258	0.9860	5e-05
0.072	29.6296	800	0.0071	0.9953	5e-05
0.079	33.3333	900	0.0000	1.0	5e-05
0.0147	37.0370	1000	0.0001	1.0	5e-05
0.0182	40.7407	1100	0.0000	1.0	5e-05
0.0363	44.4444	1200	0.0000	1.0	5e-05
0.0116	48.1481	1300	0.0000	1.0	5e-05
0.0101	51.8519	1400	0.0001	1.0	5e-05
0.0454	55.5556	1500	0.0280	0.9953	5e-05
0.1	59.2593	1600	0.0036	1.0	5e-05
0.0875	62.9630	1700	0.0726	0.9907	5e-05
0.165	66.6667	1800	0.0081	0.9953	5e-05
0.0136	70.3704	1900	0.0636	0.9907	5e-05
0.0541	74.0741	2000	0.0036	1.0	5e-05
0.0202	77.7778	2100	0.0000	1.0	1e-05
0.0	81.4815	2200	0.0000	1.0	1e-05
0.0	85.1852	2300	0.0000	1.0	1e-05
0.0	88.8889	2400	0.0000	1.0	1e-05
0.0	92.5926	2500	0.0000	1.0	1e-05
0.0	96.2963	2600	0.0000	1.0	1e-05
0.0	100.0	2700	0.0000	1.0	1e-05
0.0002	103.7037	2800	0.0000	1.0	1e-05
0.0001	107.4074	2900	0.0000	1.0	1e-05
0.0	111.1111	3000	0.0000	1.0	1e-05
0.0	114.8148	3100	0.0000	1.0	1e-05
0.0	118.5185	3200	0.0000	1.0	1e-05
0.0	122.2222	3300	0.0000	1.0	1e-05
0.0	125.9259	3400	0.0000	1.0	1e-05
0.0	129.6296	3500	0.0000	1.0	1e-05
0.0	133.3333	3600	0.0000	1.0	1e-05
0.0	137.0370	3700	0.0000	1.0	1e-05
0.0	140.7407	3800	0.0000	1.0	1e-05
0.0	144.4444	3900	0.0000	1.0	1e-05
0.0	148.1481	4000	0.0000	1.0	1e-05

Framework versions

• Transformers 4.48.1
• Pytorch 2.3.1.post300
• Datasets 3.2.0
• Tokenizers 0.21.0