scenario-kd-scr-ner-full-xlmr_data-univner_full44

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

• Loss: 158.9407
• Precision: 0.5487
• Recall: 0.5432
• F1: 0.5459
• Accuracy: 0.9588

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: 44
• 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
460.9453	0.2911	500	385.8351	0.0	0.0	0.0	0.9241
363.4557	0.5822	1000	346.2052	0.4279	0.0544	0.0965	0.9263
332.2012	0.8732	1500	325.9525	0.3466	0.0752	0.1235	0.9270
311.0392	1.1643	2000	305.0690	0.2719	0.1538	0.1965	0.9302
293.6181	1.4554	2500	289.0764	0.2902	0.1616	0.2076	0.9319
278.0759	1.7465	3000	274.3208	0.3437	0.1727	0.2299	0.9351
263.7235	2.0375	3500	261.5710	0.3720	0.2477	0.2974	0.9395
250.8804	2.3286	4000	251.6802	0.4335	0.2371	0.3065	0.9399
241.6486	2.6197	4500	241.4983	0.3763	0.3122	0.3413	0.9431
232.1585	2.9108	5000	234.7022	0.4332	0.2502	0.3172	0.9419
223.428	3.2019	5500	225.8982	0.4283	0.3369	0.3771	0.9458
215.7432	3.4929	6000	218.6250	0.4172	0.3455	0.3780	0.9469
208.5821	3.7840	6500	212.5754	0.4244	0.4311	0.4277	0.9480
202.7347	4.0751	7000	206.2203	0.4440	0.4046	0.4233	0.9502
196.2593	4.3662	7500	200.8361	0.4877	0.4275	0.4556	0.9518
191.1748	4.6573	8000	196.1823	0.4735	0.4281	0.4496	0.9524
186.3328	4.9483	8500	191.5347	0.4679	0.4555	0.4616	0.9531
180.9869	5.2394	9000	187.5859	0.4850	0.4979	0.4914	0.9549
176.8171	5.5305	9500	183.7527	0.4858	0.5217	0.5031	0.9551
173.9635	5.8216	10000	180.3877	0.5310	0.4719	0.4997	0.9553
170.1918	6.1126	10500	177.5785	0.5234	0.4582	0.4887	0.9556
167.0458	6.4037	11000	174.8427	0.5411	0.4620	0.4984	0.9554
164.3432	6.6948	11500	171.9410	0.5348	0.5089	0.5215	0.9570
161.6384	6.9859	12000	169.9951	0.5304	0.5017	0.5156	0.9573
159.3906	7.2770	12500	167.7097	0.5450	0.5037	0.5235	0.9576
157.2167	7.5680	13000	165.9562	0.5248	0.5260	0.5254	0.9578
155.8673	7.8591	13500	164.2853	0.5485	0.5210	0.5344	0.9581
153.9819	8.1502	14000	162.9678	0.5385	0.5161	0.5271	0.9581
152.5708	8.4413	14500	161.7499	0.5424	0.5385	0.5404	0.9583
151.7945	8.7324	15000	160.7884	0.5528	0.5282	0.5402	0.9585
150.7972	9.0234	15500	160.0030	0.5441	0.5455	0.5448	0.9590
149.6132	9.3145	16000	159.5214	0.5446	0.5474	0.5460	0.9585
149.1168	9.6056	16500	159.1108	0.5540	0.5376	0.5457	0.9588
149.234	9.8967	17000	158.9407	0.5487	0.5432	0.5459	0.9588

Framework versions

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