og-deberta-extra-o

This model is a fine-tuned version of microsoft/deberta-base on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 0.5184
• Precision: 0.5981
• Recall: 0.6667
• F1: 0.6305
• Accuracy: 0.9226

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: 2e-05
• train_batch_size: 8
• eval_batch_size: 8
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 25

Training results

Training Loss	Epoch	Step	Validation Loss	Precision	Recall	F1	Accuracy
No log	1.0	55	0.4813	0.2863	0.3467	0.3136	0.8720
No log	2.0	110	0.3469	0.4456	0.4587	0.4520	0.9010
No log	3.0	165	0.3166	0.5206	0.5387	0.5295	0.9147
No log	4.0	220	0.3338	0.4899	0.584	0.5328	0.9087
No log	5.0	275	0.3166	0.5625	0.648	0.6022	0.9198
No log	6.0	330	0.3464	0.5707	0.6027	0.5863	0.9207
No log	7.0	385	0.3548	0.5489	0.6133	0.5793	0.9207
No log	8.0	440	0.4005	0.6125	0.6027	0.6075	0.9210
No log	9.0	495	0.4185	0.5763	0.6347	0.6041	0.9171
0.2019	10.0	550	0.4174	0.5596	0.6507	0.6017	0.9179
0.2019	11.0	605	0.4558	0.5603	0.632	0.5940	0.9179
0.2019	12.0	660	0.4615	0.5632	0.6533	0.6049	0.9166
0.2019	13.0	715	0.4899	0.5815	0.6187	0.5995	0.9208
0.2019	14.0	770	0.4800	0.5581	0.64	0.5963	0.9186
0.2019	15.0	825	0.4752	0.5905	0.6613	0.6239	0.9212
0.2019	16.0	880	0.5014	0.5773	0.6373	0.6058	0.9174
0.2019	17.0	935	0.5095	0.5917	0.6453	0.6173	0.9195
0.2019	18.0	990	0.5249	0.5807	0.6427	0.6101	0.9203
0.0077	19.0	1045	0.5086	0.5761	0.656	0.6135	0.9222
0.0077	20.0	1100	0.5108	0.5962	0.6693	0.6307	0.9219
0.0077	21.0	1155	0.5144	0.5977	0.6853	0.6385	0.9231
0.0077	22.0	1210	0.5176	0.5990	0.6613	0.6286	0.9229
0.0077	23.0	1265	0.5171	0.6039	0.6667	0.6337	0.9226
0.0077	24.0	1320	0.5184	0.6043	0.672	0.6364	0.9226
0.0077	25.0	1375	0.5184	0.5981	0.6667	0.6305	0.9226

Framework versions

• Transformers 4.23.1
• Pytorch 1.12.1+cu113
• Datasets 2.6.1
• Tokenizers 0.13.1