doc2txt/tst_lilt_xlm_ft

test

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

• Loss: 1.6516
• Precision: 0.7245
• Recall: 0.7621
• F1: 0.7428
• Accuracy: 0.7700

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: 8
• eval_batch_size: 8
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 30

Training results

Training Loss	Epoch	Step	Validation Loss	Precision	Recall	F1	Accuracy
No log	1.33	100	0.9064	0.4989	0.6694	0.5717	0.6558
No log	2.67	200	0.9830	0.5946	0.5986	0.5966	0.6988
No log	4.0	300	0.8347	0.6432	0.6943	0.6678	0.7418
No log	5.33	400	0.8003	0.6759	0.7341	0.7038	0.7710
0.6429	6.67	500	0.9784	0.6887	0.7336	0.7104	0.7645
0.6429	8.0	600	0.9918	0.7099	0.7529	0.7308	0.7565
0.6429	9.33	700	1.1164	0.7102	0.7264	0.7182	0.7528
0.6429	10.67	800	1.3786	0.6997	0.7621	0.7296	0.7429
0.6429	12.0	900	1.2818	0.7168	0.7529	0.7344	0.7617
0.106	13.33	1000	1.3933	0.7004	0.7407	0.7200	0.7465
0.106	14.67	1100	1.3226	0.7000	0.7641	0.7306	0.7653
0.106	16.0	1200	1.5013	0.7166	0.7509	0.7333	0.7508
0.106	17.33	1300	1.4213	0.7165	0.7427	0.7294	0.7732
0.106	18.67	1400	1.4495	0.7144	0.7366	0.7254	0.7722
0.0248	20.0	1500	1.5319	0.7226	0.7326	0.7275	0.7717
0.0248	21.33	1600	1.5563	0.7232	0.7626	0.7424	0.7731
0.0248	22.67	1700	1.5967	0.7364	0.7657	0.7507	0.7734
0.0248	24.0	1800	1.5916	0.7375	0.7616	0.7494	0.7773
0.0248	25.33	1900	1.6402	0.7267	0.7504	0.7383	0.7719
0.0069	26.67	2000	1.6516	0.7250	0.7575	0.7409	0.7688
0.0069	28.0	2100	1.6539	0.7262	0.7621	0.7437	0.7697
0.0069	29.33	2200	1.6516	0.7245	0.7621	0.7428	0.7700

Framework versions

• Transformers 4.35.2
• Pytorch 2.1.0+cu121
• Datasets 2.16.1
• Tokenizers 0.15.1