mehdie/fine_tuned_XLM-roberta

fine_tuned_roberta

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

• Loss: 0.3144
• F1: 0.44
• F5: 0.5013
• Precision: 0.3333
• Recall: 0.6471

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: 256
• eval_batch_size: 256
• seed: 42
• distributed_type: multi-GPU
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 50
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	F1	F5	Precision	Recall
No log	1.0	9	0.3437	0.0	0.0	0.0	0.0
No log	2.0	18	0.3128	0.0	0.0	0.0	0.0
No log	3.0	27	0.3089	0.0	0.0	0.0	0.0
No log	4.0	36	0.3058	0.0	0.0	0.0	0.0
No log	5.0	45	0.2960	0.0	0.0	0.0	0.0
No log	6.0	54	0.2744	0.0	0.0	0.0	0.0
No log	7.0	63	0.2674	0.0	0.0	0.0	0.0
No log	8.0	72	0.2744	0.0	0.0	0.0	0.0
No log	9.0	81	0.2745	0.0	0.0	0.0	0.0
No log	10.0	90	0.2849	0.0	0.0	0.0	0.0
No log	11.0	99	0.3079	0.0	0.0	0.0	0.0
No log	12.0	108	0.2686	0.0	0.0	0.0	0.0
No log	13.0	117	0.2856	0.0	0.0	0.0	0.0
No log	14.0	126	0.3047	0.0	0.0	0.0	0.0
No log	15.0	135	0.2697	0.0	0.0	0.0	0.0
No log	16.0	144	0.2783	0.0	0.0	0.0	0.0
No log	17.0	153	0.2816	0.0	0.0	0.0	0.0
No log	18.0	162	0.2660	0.0	0.0	0.0	0.0
No log	19.0	171	0.3168	0.0	0.0	0.0	0.0
No log	20.0	180	0.2796	0.0	0.0	0.0	0.0
No log	21.0	189	0.2956	0.0	0.0	0.0	0.0
No log	22.0	198	0.2610	0.0	0.0	0.0	0.0
No log	23.0	207	0.2680	0.0	0.0	0.0	0.0
No log	24.0	216	0.2836	0.0	0.0	0.0	0.0
No log	25.0	225	0.2841	0.0588	0.0499	0.1111	0.04
No log	26.0	234	0.3023	0.2128	0.2077	0.2273	0.2
No log	27.0	243	0.3097	0.2903	0.3135	0.2432	0.36
No log	28.0	252	0.3118	0.2963	0.3049	0.2759	0.32
No log	29.0	261	0.3110	0.3256	0.3065	0.3889	0.28
No log	30.0	270	0.3261	0.2927	0.2700	0.375	0.24
No log	31.0	279	0.3167	0.4	0.4144	0.3667	0.44
No log	32.0	288	0.2997	0.2857	0.2663	0.3529	0.24
No log	33.0	297	0.3202	0.3385	0.3712	0.275	0.44
No log	34.0	306	0.2841	0.3922	0.3951	0.3846	0.4
No log	35.0	315	0.3986	0.4048	0.4788	0.2881	0.68
No log	36.0	324	0.2881	0.4082	0.4050	0.4167	0.4
No log	37.0	333	0.2719	0.3500	0.3195	0.4667	0.28
No log	38.0	342	0.3239	0.4615	0.5062	0.375	0.6
No log	39.0	351	0.2652	0.4906	0.5014	0.4643	0.52
No log	40.0	360	0.2813	0.5	0.5213	0.4516	0.56
No log	41.0	369	0.3664	0.4507	0.5081	0.3478	0.64
No log	42.0	378	0.2577	0.4651	0.4379	0.5556	0.4
No log	43.0	387	0.3193	0.4928	0.5507	0.3864	0.68
No log	44.0	396	0.2627	0.4615	0.4684	0.4444	0.48
No log	45.0	405	0.2850	0.4643	0.4841	0.4194	0.52
No log	46.0	414	0.2971	0.5574	0.5986	0.4722	0.68
No log	47.0	423	0.2804	0.5185	0.5336	0.4828	0.56
No log	48.0	432	0.2845	0.5091	0.5274	0.4667	0.56
No log	49.0	441	0.3005	0.5312	0.5797	0.4359	0.68
No log	50.0	450	0.2981	0.5079	0.5514	0.4211	0.64

Framework versions

• Transformers 4.38.2
• Pytorch 2.3.0a0+ebedce2
• Datasets 2.17.1
• Tokenizers 0.15.2