distilbert-classn-LinearAlg-finetuned-pred-span-width-2

This model is a fine-tuned version of dslim/distilbert-NER on the None dataset. It achieves the following results on the evaluation set:

• Loss: 0.6166
• Accuracy: 0.8333
• F1: 0.8321
• Precision: 0.8444
• Recall: 0.8333

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: 2
• eval_batch_size: 2
• seed: 42
• gradient_accumulation_steps: 2
• total_train_batch_size: 4
• optimizer: Use OptimizerNames.ADAMW_TORCH with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
• lr_scheduler_type: linear
• lr_scheduler_warmup_steps: 500
• num_epochs: 20
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy	F1	Precision	Recall
5.0405	0.6849	50	2.4804	0.0556	0.0339	0.0253	0.0556
4.9397	1.3699	100	2.4514	0.0714	0.0438	0.0322	0.0714
4.8593	2.0548	150	2.4005	0.0794	0.0627	0.0607	0.0794
4.7519	2.7397	200	2.3404	0.1905	0.1699	0.1849	0.1905
4.7117	3.4247	250	2.2563	0.2698	0.2785	0.3828	0.2698
4.4979	4.1096	300	2.1144	0.3810	0.3625	0.3976	0.3810
4.1155	4.7945	350	1.9127	0.5635	0.5591	0.6445	0.5635
3.5796	5.4795	400	1.6802	0.6032	0.6019	0.7030	0.6032
3.0998	6.1644	450	1.4151	0.6984	0.6879	0.7921	0.6984
2.5176	6.8493	500	1.1687	0.7698	0.7665	0.7902	0.7698
1.9411	7.5342	550	0.9922	0.7619	0.7633	0.8083	0.7619
1.4025	8.2192	600	0.8374	0.8095	0.8088	0.8457	0.8095
1.0761	8.9041	650	0.7305	0.8175	0.8124	0.8412	0.8175
0.8084	9.5890	700	0.6920	0.8254	0.8202	0.8502	0.8254
0.5516	10.2740	750	0.6456	0.8333	0.8328	0.8705	0.8333
0.4201	10.9589	800	0.6497	0.8175	0.8102	0.8566	0.8175
0.2738	11.6438	850	0.5939	0.8333	0.8337	0.8524	0.8333
0.235	12.3288	900	0.6067	0.8413	0.8397	0.8641	0.8413
0.1387	13.0137	950	0.5975	0.8333	0.8306	0.8496	0.8333
0.1154	13.6986	1000	0.5704	0.8413	0.8389	0.8515	0.8413
0.0715	14.3836	1050	0.5859	0.8413	0.8397	0.8536	0.8413
0.0741	15.0685	1100	0.5732	0.8413	0.8393	0.8510	0.8413
0.0545	15.7534	1150	0.6005	0.8333	0.8310	0.8512	0.8333
0.0354	16.4384	1200	0.6069	0.8413	0.8398	0.8564	0.8413
0.0435	17.1233	1250	0.6056	0.8413	0.8389	0.8515	0.8413
0.0305	17.8082	1300	0.6066	0.8413	0.8393	0.8558	0.8413
0.02	18.4932	1350	0.6091	0.8333	0.8315	0.8427	0.8333
0.0271	19.1781	1400	0.6121	0.8333	0.8315	0.8427	0.8333
0.0211	19.8630	1450	0.6166	0.8333	0.8321	0.8444	0.8333

Framework versions

• Transformers 4.48.3
• Pytorch 2.5.1+cu124
• Datasets 3.3.1
• Tokenizers 0.21.0