scenario-kd-scr-ner-full-mdeberta_data-univner_full55

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: 182.5497
• Precision: 0.6804
• Recall: 0.6154
• F1: 0.6463
• Accuracy: 0.9657

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
627.6279	0.2911	500	560.3030	0.0	0.0	0.0	0.9241
529.0723	0.5822	1000	503.1468	0.3145	0.0378	0.0675	0.9253
481.2237	0.8732	1500	462.0725	0.3110	0.0811	0.1286	0.9284
443.6023	1.1643	2000	431.7476	0.4241	0.0822	0.1378	0.9304
413.7332	1.4554	2500	404.1758	0.4897	0.3199	0.3870	0.9448
389.6206	1.7465	3000	381.5862	0.5329	0.3800	0.4437	0.9494
368.3142	2.0375	3500	363.5359	0.5888	0.3769	0.4596	0.9507
349.4665	2.3286	4000	346.6397	0.5410	0.4793	0.5083	0.9539
333.3893	2.6197	4500	331.5422	0.6223	0.4291	0.5079	0.9550
318.8641	2.9108	5000	316.8669	0.5984	0.5612	0.5792	0.9597
303.8825	3.2019	5500	303.3675	0.6190	0.5569	0.5863	0.9608
290.8802	3.4929	6000	291.3924	0.6347	0.5390	0.5830	0.9606
279.9562	3.7840	6500	281.3740	0.6484	0.5403	0.5894	0.9613
268.853	4.0751	7000	270.4638	0.6513	0.5578	0.6009	0.9615
257.9733	4.3662	7500	260.5476	0.6536	0.5817	0.6156	0.9635
248.9305	4.6573	8000	251.8452	0.6631	0.5926	0.6258	0.9638
240.7242	4.9483	8500	243.8925	0.6587	0.5882	0.6215	0.9633
232.3709	5.2394	9000	236.3189	0.6514	0.6077	0.6288	0.9640
224.6698	5.5305	9500	229.3991	0.6675	0.5722	0.6162	0.9629
218.3664	5.8216	10000	223.3077	0.6788	0.5823	0.6269	0.9639
212.9249	6.1126	10500	217.2704	0.6717	0.6003	0.6340	0.9643
206.6058	6.4037	11000	211.8754	0.6570	0.6226	0.6393	0.9649
201.722	6.6948	11500	207.1151	0.6680	0.6210	0.6436	0.9650
197.034	6.9859	12000	202.9470	0.6805	0.6047	0.6403	0.9649
192.5555	7.2770	12500	199.1373	0.6749	0.6130	0.6425	0.9651
189.1607	7.5680	13000	195.9332	0.6605	0.6279	0.6438	0.9652
186.1884	7.8591	13500	193.1577	0.6772	0.6057	0.6395	0.9652
183.2947	8.1502	14000	190.2176	0.6697	0.6318	0.6502	0.9654
180.5764	8.4413	14500	187.9859	0.6970	0.6091	0.6501	0.9657
178.5341	8.7324	15000	186.4189	0.6843	0.5976	0.6380	0.9645
176.79	9.0234	15500	184.5720	0.6846	0.6198	0.6506	0.9661
175.528	9.3145	16000	183.8221	0.7059	0.5905	0.6431	0.9650
174.3179	9.6056	16500	182.7365	0.6842	0.6188	0.6498	0.9658
174.2736	9.8967	17000	182.5497	0.6804	0.6154	0.6463	0.9657

Framework versions

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