scenario-TCR_data-en-massive_all_1_1

This model is a fine-tuned version of microsoft/mdeberta-v3-base on the massive dataset. It achieves the following results on the evaluation set:

• Loss: 2.6335
• Accuracy: 0.7257
• F1: 0.6761

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: 5e-05
• train_batch_size: 32
• eval_batch_size: 32
• seed: 66
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 30

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy	F1
No log	0.28	100	2.9382	0.2614	0.0710
No log	0.56	200	1.9636	0.5368	0.2848
No log	0.83	300	1.7094	0.5934	0.3887
No log	1.11	400	1.5733	0.6305	0.4633
1.8822	1.39	500	1.4046	0.6635	0.5200
1.8822	1.67	600	1.4016	0.6794	0.5558
1.8822	1.94	700	1.4019	0.6775	0.5858
1.8822	2.22	800	1.3179	0.7026	0.6044
1.8822	2.5	900	1.3087	0.7145	0.6295
0.576	2.78	1000	1.4452	0.6947	0.6119
0.576	3.06	1100	1.5017	0.6958	0.6297
0.576	3.33	1200	1.3701	0.7107	0.6439
0.576	3.61	1300	1.4868	0.7064	0.6435
0.576	3.89	1400	1.3839	0.7175	0.6397
0.3185	4.17	1500	1.5691	0.7013	0.6411
0.3185	4.44	1600	1.5106	0.7084	0.6481
0.3185	4.72	1700	1.6129	0.6979	0.6499
0.3185	5.0	1800	1.5121	0.7142	0.6551
0.3185	5.28	1900	1.6968	0.7039	0.6432
0.1966	5.56	2000	1.7057	0.7012	0.6333
0.1966	5.83	2100	1.6411	0.7165	0.6564
0.1966	6.11	2200	1.5510	0.7274	0.6709
0.1966	6.39	2300	1.7691	0.7172	0.6623
0.1966	6.67	2400	1.7955	0.7152	0.6529
0.156	6.94	2500	1.9122	0.7018	0.6548
0.156	7.22	2600	1.7143	0.7242	0.6694
0.156	7.5	2700	1.9184	0.7071	0.6528
0.156	7.78	2800	1.9581	0.7086	0.6454
0.156	8.06	2900	1.7750	0.7203	0.6643
0.0983	8.33	3000	1.9790	0.7136	0.6658
0.0983	8.61	3100	1.9127	0.7101	0.6499
0.0983	8.89	3200	2.0017	0.7121	0.6501
0.0983	9.17	3300	1.9420	0.7216	0.6650
0.0983	9.44	3400	2.0679	0.7082	0.6517
0.0767	9.72	3500	2.1093	0.7046	0.6458
0.0767	10.0	3600	2.1402	0.7126	0.6600
0.0767	10.28	3700	2.0547	0.7157	0.6578
0.0767	10.56	3800	2.1029	0.7180	0.6624
0.0767	10.83	3900	2.2774	0.7075	0.6501
0.0532	11.11	4000	2.2711	0.7005	0.6460
0.0532	11.39	4100	2.2347	0.7038	0.6500
0.0532	11.67	4200	2.3489	0.6997	0.6462
0.0532	11.94	4300	2.3262	0.7092	0.6539
0.0532	12.22	4400	2.4171	0.6990	0.6523
0.0378	12.5	4500	2.2400	0.7145	0.6600
0.0378	12.78	4600	2.2622	0.7107	0.6518
0.0378	13.06	4700	2.2886	0.6952	0.6397
0.0378	13.33	4800	2.2268	0.7128	0.6570
0.0378	13.61	4900	2.3858	0.7022	0.6453
0.0307	13.89	5000	2.2298	0.7171	0.6609
0.0307	14.17	5100	2.3298	0.7183	0.6599
0.0307	14.44	5200	2.3642	0.7117	0.6502
0.0307	14.72	5300	2.4279	0.7179	0.6681
0.0307	15.0	5400	2.5524	0.6995	0.6481
0.0264	15.28	5500	2.4293	0.7121	0.6596
0.0264	15.56	5600	2.3810	0.7163	0.6583
0.0264	15.83	5700	2.2901	0.7317	0.6745
0.0264	16.11	5800	2.3646	0.7250	0.6696
0.0264	16.39	5900	2.3795	0.7233	0.6718
0.019	16.67	6000	2.5199	0.7153	0.6647
0.019	16.94	6100	2.4350	0.7222	0.6719
0.019	17.22	6200	2.4837	0.7180	0.6702
0.019	17.5	6300	2.4684	0.7230	0.6756
0.019	17.78	6400	2.4124	0.7241	0.6743
0.0144	18.06	6500	2.5430	0.7170	0.6709
0.0144	18.33	6600	2.5298	0.7104	0.6599
0.0144	18.61	6700	2.4784	0.7217	0.6716
0.0144	18.89	6800	2.5899	0.7101	0.6703
0.0144	19.17	6900	2.4036	0.7317	0.6815
0.0127	19.44	7000	2.5389	0.7188	0.6696
0.0127	19.72	7100	2.4397	0.7263	0.6767
0.0127	20.0	7200	2.3838	0.7264	0.6734
0.0127	20.28	7300	2.4933	0.7222	0.6763
0.0127	20.56	7400	2.4831	0.7291	0.6773
0.0077	20.83	7500	2.4833	0.7255	0.6747
0.0077	21.11	7600	2.5969	0.7188	0.6728
0.0077	21.39	7700	2.5866	0.7180	0.6739
0.0077	21.67	7800	2.5581	0.7255	0.6799
0.0077	21.94	7900	2.5420	0.7266	0.6764
0.0052	22.22	8000	2.6534	0.7184	0.6670
0.0052	22.5	8100	2.5060	0.7286	0.6797
0.0052	22.78	8200	2.5219	0.7283	0.6823
0.0052	23.06	8300	2.5787	0.7220	0.6804
0.0052	23.33	8400	2.6081	0.7228	0.6784
0.0047	23.61	8500	2.5537	0.7271	0.6786
0.0047	23.89	8600	2.6520	0.7229	0.6776
0.0047	24.17	8700	2.6277	0.7261	0.6791
0.0047	24.44	8800	2.6475	0.7231	0.6759
0.0047	24.72	8900	2.6349	0.7232	0.6754
0.0031	25.0	9000	2.5821	0.7256	0.6747
0.0031	25.28	9100	2.6122	0.7241	0.6744
0.0031	25.56	9200	2.6335	0.7223	0.6727
0.0031	25.83	9300	2.6440	0.7237	0.6736
0.0031	26.11	9400	2.6027	0.7257	0.6746
0.0017	26.39	9500	2.6251	0.7240	0.6735
0.0017	26.67	9600	2.7213	0.7177	0.6711
0.0017	26.94	9700	2.7145	0.7190	0.6712
0.0017	27.22	9800	2.6901	0.7208	0.6722
0.0017	27.5	9900	2.6853	0.7207	0.6724
0.0015	27.78	10000	2.6557	0.7223	0.6731
0.0015	28.06	10100	2.6671	0.7224	0.6728
0.0015	28.33	10200	2.6418	0.7236	0.6744
0.0015	28.61	10300	2.6298	0.7255	0.6755
0.0015	28.89	10400	2.6226	0.7265	0.6775
0.0008	29.17	10500	2.6252	0.7267	0.6773
0.0008	29.44	10600	2.6322	0.7262	0.6766
0.0008	29.72	10700	2.6345	0.7255	0.6761
0.0008	30.0	10800	2.6335	0.7257	0.6761

Framework versions

• Transformers 4.33.3
• Pytorch 2.1.1+cu121
• Datasets 2.14.5
• Tokenizers 0.13.3