scenario-kd-scr-ner-full-xlmr_data-univner_half55

This model is a fine-tuned version of haryoaw/scenario-TCR-NER_data-univner_half on the None dataset. It achieves the following results on the evaluation set:

• Loss: 239.5321
• Precision: 0.3634
• Recall: 0.2698
• F1: 0.3097
• Accuracy: 0.9371

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: 8
• eval_batch_size: 32
• seed: 55
• gradient_accumulation_steps: 4
• total_train_batch_size: 32
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 10

Training results

Training Loss	Epoch	Step	Validation Loss	Precision	Recall	F1	Accuracy
443.8944	0.5828	500	368.5827	1.0	0.0003	0.0006	0.9241
344.6514	1.1655	1000	338.3108	0.4198	0.0238	0.0451	0.9249
317.7911	1.7483	1500	323.1518	0.3373	0.0781	0.1268	0.9266
295.7283	2.3310	2000	304.1432	0.3776	0.0879	0.1426	0.9282
279.1692	2.9138	2500	298.1003	0.3030	0.1619	0.2110	0.9301
265.46	3.4965	3000	283.4411	0.3299	0.1756	0.2292	0.9326
253.3522	4.0793	3500	276.4803	0.3419	0.1991	0.2517	0.9335
243.6295	4.6620	4000	268.1132	0.3623	0.2144	0.2694	0.9355
235.7751	5.2448	4500	260.5050	0.3808	0.1952	0.2581	0.9358
229.31	5.8275	5000	255.4243	0.3822	0.2135	0.2740	0.9358
222.7415	6.4103	5500	253.6783	0.3210	0.2489	0.2804	0.9345
218.7321	6.9930	6000	250.1186	0.3372	0.2663	0.2976	0.9354
213.8638	7.5758	6500	245.7943	0.3533	0.2519	0.2941	0.9362
211.1232	8.1585	7000	241.6974	0.3942	0.2450	0.3022	0.9382
208.2374	8.7413	7500	241.2330	0.3854	0.2630	0.3127	0.9375
206.2932	9.3240	8000	240.2229	0.3769	0.2672	0.3127	0.9373
205.5458	9.9068	8500	239.5321	0.3634	0.2698	0.3097	0.9371

Framework versions

• Transformers 4.44.2
• Pytorch 2.1.1+cu121
• Datasets 2.14.5
• Tokenizers 0.19.1