scenario-kd-scr-ner-full-xlmr_data-univner_half66

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: 239.7216
• Precision: 0.3760
• Recall: 0.2925
• F1: 0.3290
• Accuracy: 0.9383

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: 66
• 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
445.7648	0.5828	500	369.2686	0.0	0.0	0.0	0.9241
345.7662	1.1655	1000	344.5894	0.3114	0.0688	0.1127	0.9255
317.4014	1.7483	1500	321.3392	0.3477	0.0814	0.1319	0.9268
296.2663	2.3310	2000	304.3693	0.4262	0.0837	0.1399	0.9278
278.6608	2.9138	2500	293.9949	0.3311	0.1453	0.2020	0.9306
265.0553	3.4965	3000	282.5093	0.3451	0.1672	0.2253	0.9328
253.4131	4.0793	3500	275.8793	0.3284	0.2137	0.2589	0.9337
243.5783	4.6620	4000	268.4692	0.3237	0.2345	0.2719	0.9341
235.5956	5.2448	4500	262.3924	0.3467	0.2506	0.2909	0.9355
228.406	5.8275	5000	257.0986	0.3634	0.2381	0.2877	0.9361
222.6923	6.4103	5500	250.0558	0.3799	0.2350	0.2904	0.9380
218.056	6.9930	6000	246.7546	0.3904	0.2479	0.3032	0.9383
213.6749	7.5758	6500	245.7390	0.3713	0.2721	0.3141	0.9373
210.549	8.1585	7000	242.9818	0.3644	0.2611	0.3043	0.9367
207.9283	8.7413	7500	240.6766	0.3721	0.2789	0.3188	0.9376
206.1746	9.3240	8000	239.1118	0.3990	0.2772	0.3271	0.9391
205.302	9.9068	8500	239.7216	0.3760	0.2925	0.3290	0.9383

Framework versions

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