kavg/LiLT-RE-FR-SIN

checkpoints

This model is a fine-tuned version of kavg/LiLT-RE-FR on the xfun dataset. It achieves the following results on the evaluation set:

• Precision: 0.3604
• Recall: 0.5707
• F1: 0.4418
• Loss: 0.2693

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: 1e-05
• train_batch_size: 8
• eval_batch_size: 2
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• lr_scheduler_warmup_ratio: 0.1
• training_steps: 10000

Training results

Training Loss	Epoch	Step	Precision	Recall	F1	Validation Loss
0.0868	41.67	500	0.3688	0.2803	0.3185	0.1679
0.0557	83.33	1000	0.3604	0.5707	0.4418	0.2693
0.0513	125.0	1500	0.3962	0.5833	0.4719	0.3008
0.0248	166.67	2000	0.4043	0.6237	0.4906	0.4857
0.0139	208.33	2500	0.4296	0.6010	0.5011	0.4227
0.004	250.0	3000	0.4177	0.6212	0.4995	0.5369
0.0084	291.67	3500	0.4255	0.6490	0.514	0.5332
0.0067	333.33	4000	0.4259	0.6389	0.5111	0.4978
0.0008	375.0	4500	0.4189	0.6263	0.5020	0.4567
0.0116	416.67	5000	0.4336	0.6515	0.5207	0.5514
0.0007	458.33	5500	0.4394	0.6414	0.5216	0.5703
0.0004	500.0	6000	0.4504	0.6540	0.5335	0.6107
0.002	541.67	6500	0.4480	0.6414	0.5275	0.5859
0.0059	583.33	7000	0.4526	0.6263	0.5254	0.6033
0.0023	625.0	7500	0.4379	0.6414	0.5205	0.6440
0.0007	666.67	8000	0.4499	0.6237	0.5228	0.5594
0.003	708.33	8500	0.4393	0.6490	0.5240	0.6276
0.0001	750.0	9000	0.4410	0.6515	0.5260	0.6132
0.001	791.67	9500	0.4376	0.6288	0.5161	0.6312
0.0001	833.33	10000	0.4415	0.6389	0.5222	0.6304

Framework versions

• Transformers 4.38.2
• Pytorch 2.1.0+cu121
• Datasets 2.18.0
• Tokenizers 0.15.1