UNER_subword_tk_en_lora_alpha_1024_drop_0.3_rank_512_seed_42

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

• Loss: 0.0633
• Precision: 0.7732
• Recall: 0.8292
• F1: 0.8002
• Accuracy: 0.9844

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: 32
• eval_batch_size: 32
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 35.0

Training results

Training Loss	Epoch	Step	Validation Loss	Precision	Recall	F1	Accuracy
No log	1.0	392	0.1362	0.2922	0.3903	0.3342	0.9569
0.2046	2.0	784	0.0889	0.5868	0.6822	0.6309	0.9745
0.085	3.0	1176	0.0772	0.6687	0.7940	0.7260	0.9778
0.0591	4.0	1568	0.0692	0.7085	0.7950	0.7493	0.9802
0.0591	5.0	1960	0.0692	0.6894	0.8251	0.7512	0.9791
0.0496	6.0	2352	0.0664	0.6937	0.8157	0.7498	0.9791
0.0448	7.0	2744	0.0671	0.7007	0.8313	0.7604	0.9797
0.0409	8.0	3136	0.0674	0.7200	0.8147	0.7644	0.9814
0.0388	9.0	3528	0.0635	0.7306	0.8478	0.7849	0.9816
0.0388	10.0	3920	0.0620	0.7481	0.8209	0.7828	0.9832
0.0357	11.0	4312	0.0586	0.7758	0.8240	0.7992	0.9844
0.0333	12.0	4704	0.0611	0.7606	0.8354	0.7963	0.9840
0.0323	13.0	5096	0.0601	0.7819	0.8240	0.8024	0.9844
0.0323	14.0	5488	0.0638	0.7203	0.8292	0.7709	0.9812
0.0303	15.0	5880	0.0600	0.7737	0.8354	0.8034	0.9841
0.0293	16.0	6272	0.0602	0.7703	0.8333	0.8006	0.9841
0.0271	17.0	6664	0.0609	0.7634	0.8416	0.8006	0.9841
0.0269	18.0	7056	0.0641	0.7569	0.8478	0.7998	0.9835
0.0269	19.0	7448	0.0594	0.7793	0.8261	0.8020	0.9849
0.0263	20.0	7840	0.0608	0.7873	0.8199	0.8032	0.9850
0.025	21.0	8232	0.0606	0.7812	0.8240	0.8020	0.9850
0.0236	22.0	8624	0.0639	0.7558	0.8364	0.7941	0.9839
0.0228	23.0	9016	0.0620	0.7668	0.8375	0.8006	0.9845
0.0228	24.0	9408	0.0612	0.7647	0.8344	0.7980	0.9842
0.0229	25.0	9800	0.0618	0.7584	0.8385	0.7965	0.9839
0.0227	26.0	10192	0.0631	0.7678	0.8385	0.8016	0.9842
0.0216	27.0	10584	0.0628	0.7883	0.8364	0.8117	0.9850
0.0216	28.0	10976	0.0611	0.7765	0.8344	0.8044	0.9849
0.0203	29.0	11368	0.0615	0.7755	0.8406	0.8068	0.9847
0.02	30.0	11760	0.0629	0.7743	0.8344	0.8032	0.9847
0.0197	31.0	12152	0.0620	0.7763	0.8333	0.8038	0.9843
0.0197	32.0	12544	0.0633	0.7750	0.8271	0.8002	0.9845
0.0197	33.0	12936	0.0631	0.7813	0.8323	0.8060	0.9845
0.0192	34.0	13328	0.0629	0.7768	0.8323	0.8036	0.9845
0.0188	35.0	13720	0.0633	0.7732	0.8292	0.8002	0.9844

Framework versions

• Transformers 4.41.1
• Pytorch 2.3.0+cu121
• Datasets 2.18.0
• Tokenizers 0.19.1