t5-small_winobias_finetuned

This model is a fine-tuned version of t5-small on the None dataset. It achieves the following results on the evaluation set:

Loss: 0.2473
Accuracy: 0.5278
Tp: 0.5
Tn: 0.0278
Fp: 0.4722
Fn: 0.0

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: 0.0003
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: 30

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy	Tp	Tn	Fp
0.6334	0.8	20	0.3622	0.5	0.5	0.0	0.5
0.4058	1.6	40	0.3510	0.5	0.5	0.0	0.5
0.3923	2.4	60	0.3511	0.5	0.5	0.0	0.5
0.376	3.2	80	0.3509	0.5	0.5	0.0	0.5
0.3749	4.0	100	0.3502	0.5	0.5	0.0	0.5
0.3895	4.8	120	0.3505	0.5	0.5	0.0	0.5
0.3624	5.6	140	0.3508	0.5	0.5	0.0	0.5
0.3754	6.4	160	0.3501	0.5	0.5	0.0	0.5
0.3702	7.2	180	0.3576	0.5	0.5	0.0	0.5
0.3748	8.0	200	0.3499	0.5	0.5	0.0	0.5
0.3715	8.8	220	0.3482	0.5	0.5	0.0	0.5
0.3576	9.6	240	0.3489	0.5	0.5	0.0	0.5
0.3659	10.4	260	0.3510	0.5	0.5	0.0	0.5
0.3565	11.2	280	0.3464	0.5	0.5	0.0	0.5
0.353	12.0	300	0.3474	0.5	0.5	0.0	0.5
0.3614	12.8	320	0.3450	0.5	0.5	0.0	0.5
0.3625	13.6	340	0.3458	0.5	0.5	0.0	0.5
0.36	14.4	360	0.3494	0.5	0.5	0.0	0.5
0.3585	15.2	380	0.3435	0.5	0.5	0.0	0.5
0.3541	16.0	400	0.3431	0.5	0.5	0.0	0.5
0.3564	16.8	420	0.3414	0.5	0.5	0.0	0.5
0.3462	17.6	440	0.3413	0.5	0.5	0.0	0.5
0.3541	18.4	460	0.3382	0.5	0.5	0.0	0.5
0.3579	19.2	480	0.3399	0.5	0.5	0.0	0.5
0.3466	20.0	500	0.3317	0.5	0.5	0.0	0.5
0.3314	20.8	520	0.3303	0.5	0.5	0.0	0.5
0.33	21.6	540	0.3246	0.5	0.5	0.0	0.5
0.3279	22.4	560	0.3154	0.5	0.5	0.0	0.5
0.3234	23.2	580	0.3050	0.5	0.5	0.0	0.5
0.3193	24.0	600	0.2947	0.5	0.5	0.0	0.5
0.3086	24.8	620	0.2849	0.5013	0.5	0.0013	0.4987
0.2912	25.6	640	0.2748	0.5013	0.5	0.0013	0.4987
0.2787	26.4	660	0.2655	0.5107	0.5	0.0107	0.4893
0.2779	27.2	680	0.2581	0.5177	0.5	0.0177	0.4823
0.2697	28.0	700	0.2527	0.5170	0.5	0.0170	0.4830
0.2669	28.8	720	0.2495	0.5259	0.5	0.0259	0.4741
0.2654	29.6	740	0.2473	0.5278	0.5	0.0278	0.4722

Framework versions

Transformers 4.26.1
Pytorch 1.13.1
Datasets 2.10.1
Tokenizers 0.13.2

Loading…

Training procedure

Training hyperparameters

The following hyperparameters were used during training:

learning_rate: 0.0003

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: 30

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy	Tp	Tn	Fp
0.6334	0.8	20	0.3622	0.5	0.5	0.0	0.5
0.4058	1.6	40	0.3510	0.5	0.5	0.0	0.5
0.3923	2.4	60	0.3511	0.5	0.5	0.0	0.5
0.376	3.2	80	0.3509	0.5	0.5	0.0	0.5
0.3749	4.0	100	0.3502	0.5	0.5	0.0	0.5
0.3895	4.8	120	0.3505	0.5	0.5	0.0	0.5
0.3624	5.6	140	0.3508	0.5	0.5	0.0	0.5
0.3754	6.4	160	0.3501	0.5	0.5	0.0	0.5
0.3702	7.2	180	0.3576	0.5	0.5	0.0	0.5
0.3748	8.0	200	0.3499	0.5	0.5	0.0	0.5
0.3715	8.8	220	0.3482	0.5	0.5	0.0	0.5
0.3576	9.6	240	0.3489	0.5	0.5	0.0	0.5
0.3659	10.4	260	0.3510	0.5	0.5	0.0	0.5
0.3565	11.2	280	0.3464	0.5	0.5	0.0	0.5
0.353	12.0	300	0.3474	0.5	0.5	0.0	0.5
0.3614	12.8	320	0.3450	0.5	0.5	0.0	0.5
0.3625	13.6	340	0.3458	0.5	0.5	0.0	0.5
0.36	14.4	360	0.3494	0.5	0.5	0.0	0.5
0.3585	15.2	380	0.3435	0.5	0.5	0.0	0.5
0.3541	16.0	400	0.3431	0.5	0.5	0.0	0.5
0.3564	16.8	420	0.3414	0.5	0.5	0.0	0.5
0.3462	17.6	440	0.3413	0.5	0.5	0.0	0.5
0.3541	18.4	460	0.3382	0.5	0.5	0.0	0.5
0.3579	19.2	480	0.3399	0.5	0.5	0.0	0.5
0.3466	20.0	500	0.3317	0.5	0.5	0.0	0.5
0.3314	20.8	520	0.3303	0.5	0.5	0.0	0.5
0.33	21.6	540	0.3246	0.5	0.5	0.0	0.5
0.3279	22.4	560	0.3154	0.5	0.5	0.0	0.5
0.3234	23.2	580	0.3050	0.5	0.5	0.0	0.5
0.3193	24.0	600	0.2947	0.5	0.5	0.0	0.5
0.3086	24.8	620	0.2849	0.5013	0.5	0.0013	0.4987
0.2912	25.6	640	0.2748	0.5013	0.5	0.0013	0.4987
0.2787	26.4	660	0.2655	0.5107	0.5	0.0107	0.4893
0.2779	27.2	680	0.2581	0.5177	0.5	0.0177	0.4823
0.2697	28.0	700	0.2527	0.5170	0.5	0.0170	0.4830
0.2669	28.8	720	0.2495	0.5259	0.5	0.0259	0.4741
0.2654	29.6	740	0.2473	0.5278	0.5	0.0278	0.4722

Framework versions

Transformers 4.26.1

Pytorch 1.13.1

Datasets 2.10.1

Tokenizers 0.13.2

henryscheible/t5-small_winobias_finetuned

t5-small_winobias_finetuned

Model description

Intended uses & limitations

Training and evaluation data

Training procedure

Training hyperparameters

Training results

Framework versions

t5-small_winobias_finetuned

Model description

Intended uses & limitations

Training and evaluation data

Training procedure

Training hyperparameters

Training results

Framework versions