scenario-MDBT-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: 2.5243
• Accuracy: 0.7361
• F1: 0.6881

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: 64
• 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	2.9524	0.3055	0.0759
No log	0.56	200	2.0666	0.4969	0.2470
No log	0.83	300	1.6852	0.5878	0.3626
No log	1.11	400	1.4751	0.6376	0.4564
1.9292	1.39	500	1.4633	0.6522	0.4887
1.9292	1.67	600	1.4477	0.6604	0.5016
1.9292	1.94	700	1.3844	0.6758	0.5746
1.9292	2.22	800	1.3272	0.6937	0.5857
1.9292	2.5	900	1.2445	0.7178	0.6192
0.6258	2.78	1000	1.3348	0.7128	0.6198
0.6258	3.06	1100	1.5354	0.6734	0.6006
0.6258	3.33	1200	1.4149	0.7001	0.6258
0.6258	3.61	1300	1.4474	0.7032	0.6405
0.6258	3.89	1400	1.5031	0.7054	0.6395
0.3592	4.17	1500	1.3733	0.7225	0.6669
0.3592	4.44	1600	1.3757	0.7317	0.6555
0.3592	4.72	1700	1.4694	0.7134	0.6539
0.3592	5.0	1800	1.4733	0.7077	0.6461
0.3592	5.28	1900	1.5077	0.7232	0.6654
0.2316	5.56	2000	1.6396	0.7069	0.6482
0.2316	5.83	2100	1.5588	0.7178	0.6599
0.2316	6.11	2200	1.5611	0.7206	0.6507
0.2316	6.39	2300	1.7029	0.7155	0.6597
0.2316	6.67	2400	1.7865	0.7048	0.6407
0.1763	6.94	2500	1.7791	0.7065	0.6555
0.1763	7.22	2600	1.8013	0.7176	0.6572
0.1763	7.5	2700	1.8034	0.7149	0.6578
0.1763	7.78	2800	2.0082	0.6872	0.6234
0.1763	8.06	2900	2.0108	0.7011	0.6388
0.1136	8.33	3000	2.0779	0.6997	0.6513
0.1136	8.61	3100	1.9805	0.7122	0.6555
0.1136	8.89	3200	2.1014	0.7010	0.6485
0.1136	9.17	3300	1.9710	0.7133	0.6537
0.1136	9.44	3400	1.9677	0.7152	0.6564
0.0964	9.72	3500	2.0902	0.7079	0.6535
0.0964	10.0	3600	2.0776	0.7083	0.6529
0.0964	10.28	3700	2.0649	0.7191	0.6647
0.0964	10.56	3800	2.0690	0.7152	0.6551
0.0964	10.83	3900	2.1585	0.7055	0.6513
0.0721	11.11	4000	2.0158	0.7236	0.6660
0.0721	11.39	4100	2.1559	0.7120	0.6616
0.0721	11.67	4200	2.0517	0.7253	0.6694
0.0721	11.94	4300	2.1721	0.7219	0.6662
0.0721	12.22	4400	2.2949	0.7079	0.6680
0.0448	12.5	4500	2.1676	0.7186	0.6685
0.0448	12.78	4600	2.0882	0.7227	0.6636
0.0448	13.06	4700	2.0149	0.7335	0.6736
0.0448	13.33	4800	2.2128	0.7243	0.6667
0.0448	13.61	4900	2.2664	0.7200	0.6577
0.0371	13.89	5000	2.3489	0.7100	0.6656
0.0371	14.17	5100	2.3454	0.7087	0.6531
0.0371	14.44	5200	2.2062	0.7296	0.6767
0.0371	14.72	5300	2.4544	0.7101	0.6661
0.0371	15.0	5400	2.2581	0.7275	0.6683
0.0227	15.28	5500	2.2904	0.7242	0.6697
0.0227	15.56	5600	2.3484	0.7152	0.6495
0.0227	15.83	5700	2.4505	0.7126	0.6599
0.0227	16.11	5800	2.2985	0.7236	0.6673
0.0227	16.39	5900	2.3929	0.7245	0.6751
0.022	16.67	6000	2.4606	0.7200	0.6643
0.022	16.94	6100	2.3481	0.7276	0.6689
0.022	17.22	6200	2.3302	0.7273	0.6724
0.022	17.5	6300	2.3566	0.7292	0.6787
0.022	17.78	6400	2.3972	0.7281	0.6785
0.0133	18.06	6500	2.5105	0.7205	0.6705
0.0133	18.33	6600	2.3785	0.7295	0.6775
0.0133	18.61	6700	2.4367	0.7220	0.6676
0.0133	18.89	6800	2.4496	0.7255	0.6690
0.0133	19.17	6900	2.4133	0.7279	0.6720
0.0097	19.44	7000	2.5588	0.7140	0.6652
0.0097	19.72	7100	2.4906	0.7210	0.6656
0.0097	20.0	7200	2.5187	0.7199	0.6619
0.0097	20.28	7300	2.4627	0.7254	0.6686
0.0097	20.56	7400	2.5543	0.7187	0.6615
0.0096	20.83	7500	2.4262	0.7259	0.6676
0.0096	21.11	7600	2.4768	0.7256	0.6699
0.0096	21.39	7700	2.5336	0.7220	0.6724
0.0096	21.67	7800	2.5221	0.7240	0.6703
0.0096	21.94	7900	2.5008	0.7269	0.6712
0.0086	22.22	8000	2.4998	0.7278	0.6703
0.0086	22.5	8100	2.4611	0.7319	0.6842
0.0086	22.78	8200	2.5119	0.7313	0.6832
0.0086	23.06	8300	2.4329	0.7300	0.6764
0.0086	23.33	8400	2.4080	0.7317	0.6822
0.007	23.61	8500	2.4054	0.7313	0.6802
0.007	23.89	8600	2.4345	0.7334	0.6851
0.007	24.17	8700	2.4735	0.7326	0.6865
0.007	24.44	8800	2.4718	0.7313	0.6843
0.007	24.72	8900	2.4391	0.7328	0.6818
0.0029	25.0	9000	2.5152	0.7290	0.6869
0.0029	25.28	9100	2.4609	0.7365	0.6908
0.0029	25.56	9200	2.4717	0.7359	0.6932
0.0029	25.83	9300	2.5283	0.7337	0.6881
0.0029	26.11	9400	2.4831	0.7342	0.6866
0.0026	26.39	9500	2.5291	0.7325	0.6861
0.0026	26.67	9600	2.5201	0.7344	0.6855
0.0026	26.94	9700	2.5496	0.7322	0.6857
0.0026	27.22	9800	2.5302	0.7332	0.6853
0.0026	27.5	9900	2.5388	0.7329	0.6871
0.0025	27.78	10000	2.5210	0.7326	0.6845
0.0025	28.06	10100	2.5482	0.7319	0.6841
0.0025	28.33	10200	2.5628	0.7315	0.6853
0.0025	28.61	10300	2.5439	0.7341	0.6870
0.0025	28.89	10400	2.5241	0.7356	0.6875
0.001	29.17	10500	2.5238	0.7354	0.6873
0.001	29.44	10600	2.5186	0.7362	0.6880
0.001	29.72	10700	2.5237	0.7360	0.6880
0.001	30.0	10800	2.5243	0.7361	0.6881

Framework versions

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