metadata

library_name: transformers
license: mit
base_model: FacebookAI/xlm-roberta-base
tags:
  - generated_from_trainer
metrics:
  - precision
  - recall
  - f1
  - accuracy
model-index:
  - name: scenario-kd-pre-ner-full_data-univner_full44
    results: []

scenario-kd-pre-ner-full_data-univner_full44

This model is a fine-tuned version of FacebookAI/xlm-roberta-base on the None dataset. It achieves the following results on the evaluation set:

Loss: 0.4381
Precision: 0.8004
Recall: 0.7801
F1: 0.7902
Accuracy: 0.9786

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: 3e-05
train_batch_size: 32
eval_batch_size: 32
seed: 44
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
1.4593	0.5828	500	0.8367	0.6935	0.6559	0.6742	0.9682
0.7232	1.1655	1000	0.7569	0.7339	0.6980	0.7155	0.9724
0.594	1.7483	1500	0.6330	0.7335	0.7451	0.7392	0.9741
0.4986	2.3310	2000	0.6003	0.7291	0.7552	0.7419	0.9746
0.446	2.9138	2500	0.5729	0.7403	0.7601	0.7501	0.9747
0.385	3.4965	3000	0.5584	0.7441	0.7617	0.7528	0.9757
0.3605	4.0793	3500	0.5602	0.7615	0.7575	0.7595	0.9758
0.3172	4.6620	4000	0.5417	0.7546	0.7725	0.7634	0.9764
0.3061	5.2448	4500	0.5329	0.7884	0.7485	0.7680	0.9769
0.2856	5.8275	5000	0.5194	0.7837	0.7618	0.7726	0.9769
0.2642	6.4103	5500	0.5154	0.7622	0.7780	0.7700	0.9765
0.2592	6.9930	6000	0.5193	0.7882	0.7572	0.7724	0.9764
0.2401	7.5758	6500	0.5123	0.7727	0.7599	0.7663	0.9763
0.2344	8.1585	7000	0.4987	0.7742	0.7736	0.7739	0.9771
0.2234	8.7413	7500	0.4914	0.7894	0.7640	0.7764	0.9777
0.2131	9.3240	8000	0.4856	0.7691	0.7827	0.7758	0.9770
0.2089	9.9068	8500	0.4898	0.7895	0.7655	0.7773	0.9773
0.1972	10.4895	9000	0.4860	0.7828	0.7726	0.7777	0.9775
0.1942	11.0723	9500	0.4787	0.7807	0.7807	0.7807	0.9776
0.1854	11.6550	10000	0.4858	0.7916	0.7635	0.7773	0.9771
0.183	12.2378	10500	0.4739	0.7924	0.7800	0.7862	0.9779
0.1781	12.8205	11000	0.4741	0.7990	0.7661	0.7822	0.9779
0.1704	13.4033	11500	0.4622	0.7937	0.7719	0.7826	0.9784
0.1698	13.9860	12000	0.4650	0.8000	0.7657	0.7825	0.9777
0.1635	14.5688	12500	0.4604	0.7913	0.7778	0.7845	0.9782
0.1605	15.1515	13000	0.4656	0.7990	0.7605	0.7793	0.9774
0.1559	15.7343	13500	0.4638	0.8001	0.7658	0.7826	0.9778
0.1531	16.3170	14000	0.4550	0.7991	0.7735	0.7861	0.9780
0.1519	16.8998	14500	0.4606	0.7949	0.7735	0.7841	0.9780
0.1482	17.4825	15000	0.4483	0.7947	0.7831	0.7889	0.9787
0.1449	18.0653	15500	0.4521	0.7947	0.7722	0.7833	0.9780
0.1407	18.6480	16000	0.4508	0.7932	0.7728	0.7829	0.9780
0.1415	19.2308	16500	0.4484	0.8031	0.7728	0.7876	0.9785
0.1385	19.8135	17000	0.4461	0.7991	0.7774	0.7881	0.9785
0.1358	20.3963	17500	0.4488	0.7970	0.7756	0.7862	0.9783
0.1358	20.9790	18000	0.4431	0.8006	0.7772	0.7887	0.9787
0.1325	21.5618	18500	0.4395	0.8053	0.7768	0.7908	0.9785
0.1322	22.1445	19000	0.4461	0.7960	0.7725	0.7841	0.9780
0.1296	22.7273	19500	0.4401	0.7988	0.7746	0.7866	0.9781
0.1288	23.3100	20000	0.4416	0.7961	0.7690	0.7823	0.9781
0.1271	23.8928	20500	0.4450	0.8024	0.7673	0.7844	0.9781
0.1246	24.4755	21000	0.4403	0.7967	0.7703	0.7833	0.9782
0.1254	25.0583	21500	0.4403	0.7976	0.7742	0.7857	0.9782
0.1231	25.6410	22000	0.4438	0.8057	0.7694	0.7872	0.9783
0.1228	26.2238	22500	0.4365	0.8058	0.7741	0.7896	0.9785
0.1224	26.8065	23000	0.4325	0.7995	0.7806	0.7899	0.9787
0.1211	27.3893	23500	0.4402	0.8058	0.7676	0.7862	0.9782
0.1202	27.9720	24000	0.4378	0.8017	0.7689	0.7849	0.9784
0.1201	28.5548	24500	0.4331	0.8000	0.7784	0.7890	0.9786
0.12	29.1375	25000	0.4317	0.7999	0.7794	0.7895	0.9787
0.1194	29.7203	25500	0.4381	0.8004	0.7801	0.7902	0.9786

Framework versions

Transformers 4.44.2
Pytorch 2.1.1+cu121
Datasets 2.14.5
Tokenizers 0.19.1