gpt2_crows_pairs_finetuned

This model is a fine-tuned version of gpt2 on the crows_pairs dataset. It achieves the following results on the evaluation set:

Loss: 2.0946
Accuracy: 0.7781
Tp: 0.3444
Tn: 0.4338
Fp: 0.1159
Fn: 0.1060

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: 64
eval_batch_size: 64
seed: 42
optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
lr_scheduler_type: linear
num_epochs: 50

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy	Tp	Tn	Fp	Fn
0.7371	1.05	20	0.7345	0.4669	0.4305	0.0364	0.5132	0.0199
0.6794	2.11	40	0.6829	0.5397	0.3013	0.2384	0.3113	0.1490
0.5972	3.16	60	0.6602	0.6291	0.3411	0.2881	0.2616	0.1093
0.4691	4.21	80	0.6568	0.6788	0.3742	0.3046	0.2450	0.0762
0.3645	5.26	100	0.5872	0.7252	0.2815	0.4437	0.1060	0.1689
0.2645	6.32	120	0.6835	0.7185	0.2318	0.4868	0.0629	0.2185
0.1698	7.37	140	0.7757	0.7483	0.2914	0.4570	0.0927	0.1589
0.1386	8.42	160	0.7445	0.7417	0.2881	0.4536	0.0960	0.1623
0.077	9.47	180	1.0591	0.7252	0.3642	0.3609	0.1887	0.0861
0.0836	10.53	200	1.0908	0.7185	0.2649	0.4536	0.0960	0.1854
0.0485	11.58	220	1.2155	0.7450	0.3709	0.3742	0.1755	0.0795
0.0298	12.63	240	1.1973	0.7417	0.3245	0.4172	0.1325	0.1258
0.0444	13.68	260	1.4213	0.7384	0.3675	0.3709	0.1788	0.0828
0.0215	14.74	280	1.4907	0.7450	0.3278	0.4172	0.1325	0.1225
0.0483	15.79	300	1.5485	0.7583	0.2781	0.4801	0.0695	0.1722
0.0129	16.84	320	1.7145	0.7550	0.2748	0.4801	0.0695	0.1755
0.0525	17.89	340	1.7827	0.7550	0.3642	0.3907	0.1589	0.0861
0.0074	18.95	360	1.6230	0.7682	0.2980	0.4702	0.0795	0.1523
0.004	20.0	380	1.8522	0.7384	0.3444	0.3940	0.1556	0.1060
0.0067	21.05	400	1.8479	0.7616	0.3046	0.4570	0.0927	0.1457
0.001	22.11	420	1.9830	0.7682	0.2947	0.4735	0.0762	0.1556
0.01	23.16	440	1.9412	0.7715	0.3113	0.4603	0.0894	0.1391
0.0048	24.21	460	2.0075	0.7649	0.3510	0.4139	0.1358	0.0993
0.0025	25.26	480	2.0912	0.7649	0.2980	0.4669	0.0828	0.1523
0.0013	26.32	500	2.1548	0.7715	0.3444	0.4272	0.1225	0.1060
0.0041	27.37	520	2.1337	0.7682	0.3543	0.4139	0.1358	0.0960
0.0005	28.42	540	2.1242	0.7550	0.3576	0.3974	0.1523	0.0927
0.0124	29.47	560	2.1297	0.7583	0.3642	0.3940	0.1556	0.0861
0.0104	30.53	580	2.0057	0.7583	0.3179	0.4404	0.1093	0.1325
0.0156	31.58	600	2.0365	0.7483	0.2881	0.4603	0.0894	0.1623
0.0003	32.63	620	1.9614	0.7649	0.3212	0.4437	0.1060	0.1291
0.0029	33.68	640	1.9658	0.7682	0.3245	0.4437	0.1060	0.1258
0.0001	34.74	660	1.9913	0.7649	0.3013	0.4636	0.0861	0.1490
0.0001	35.79	680	2.0039	0.7649	0.3013	0.4636	0.0861	0.1490
0.0004	36.84	700	1.9657	0.7715	0.3146	0.4570	0.0927	0.1358
0.0003	37.89	720	1.9787	0.7748	0.3245	0.4503	0.0993	0.1258
0.0007	38.95	740	1.9888	0.7781	0.3377	0.4404	0.1093	0.1126
0.0002	40.0	760	2.0293	0.7682	0.3477	0.4205	0.1291	0.1026
0.0002	41.05	780	1.9914	0.7781	0.3245	0.4536	0.0960	0.1258
0.0003	42.11	800	2.0444	0.7583	0.2914	0.4669	0.0828	0.1589
0.0072	43.16	820	2.0247	0.7649	0.3278	0.4371	0.1126	0.1225
0.0001	44.21	840	2.0398	0.7682	0.3278	0.4404	0.1093	0.1225
0.0001	45.26	860	2.0358	0.7682	0.3278	0.4404	0.1093	0.1225
0.0011	46.32	880	2.0432	0.7682	0.3278	0.4404	0.1093	0.1225
0.0001	47.37	900	2.0923	0.7781	0.3444	0.4338	0.1159	0.1060
0.0	48.42	920	2.0975	0.7781	0.3444	0.4338	0.1159	0.1060
0.0002	49.47	940	2.0946	0.7781	0.3444	0.4338	0.1159	0.1060

Framework versions

Transformers 4.26.1
Pytorch 1.13.1
Datasets 2.10.1
Tokenizers 0.13.2

henryscheible/gpt2_crows_pairs_finetuned

gpt2_crows_pairs_finetuned

Model description

Intended uses & limitations

Training and evaluation data

Training procedure

Training hyperparameters

Training results

Framework versions

为此模型做种