kavg/LiLT-RE-PT-SIN

checkpoints

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

• Precision: 0.3631
• Recall: 0.4823
• F1: 0.4143
• Loss: 0.1671

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
0.0907	41.67	500	0.3315	0.3106	0.3207	0.2039
0.0766	83.33	1000	0.3631	0.4823	0.4143	0.1671
0.0639	125.0	1500	0.3640	0.6086	0.4556	0.2525
0.0309	166.67	2000	0.3973	0.6010	0.4784	0.2339
0.0318	208.33	2500	0.4045	0.6414	0.4961	0.3325
0.0144	250.0	3000	0.4268	0.6187	0.5052	0.3513
0.0163	291.67	3500	0.4273	0.6086	0.5021	0.2880
0.0062	333.33	4000	0.4368	0.6288	0.5155	0.3064
0.0115	375.0	4500	0.4386	0.6313	0.5176	0.3283
0.0168	416.67	5000	0.4373	0.6162	0.5115	0.3258
0.0062	458.33	5500	0.4530	0.6086	0.5194	0.3467
0.0074	500.0	6000	0.4569	0.6162	0.5247	0.3401
0.0037	541.67	6500	0.4559	0.6136	0.5231	0.3526
0.008	583.33	7000	0.4650	0.6035	0.5253	0.3076
0.0045	625.0	7500	0.4610	0.6111	0.5255	0.3799
0.0045	666.67	8000	0.4551	0.6136	0.5226	0.3692
0.0052	708.33	8500	0.4535	0.6162	0.5225	0.3492
0.0002	750.0	9000	0.4537	0.6061	0.5189	0.4075
0.0027	791.67	9500	0.4581	0.6212	0.5273	0.3816
0.0009	833.33	10000	0.4569	0.6162	0.5247	0.3834

Framework versions

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