question_classification

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

• Loss: 1.0442
• Accuracy: 0.9054
• F1: 0.9045

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: 4
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 35

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy	F1
No log	1.0	204	1.2496	0.6132	0.6098
No log	2.0	408	0.7790	0.7249	0.7231
1.2508	3.0	612	0.6412	0.8023	0.8038
1.2508	4.0	816	0.5420	0.8682	0.8681
0.318	5.0	1020	0.7027	0.8453	0.8428
0.318	6.0	1224	0.6174	0.8625	0.8629
0.318	7.0	1428	0.6363	0.8768	0.8772
0.121	8.0	1632	0.7726	0.8682	0.8695
0.121	9.0	1836	1.0105	0.8739	0.8734
0.043	10.0	2040	0.9210	0.8854	0.8855
0.043	11.0	2244	0.9544	0.8825	0.8794
0.043	12.0	2448	0.8467	0.8825	0.8825
0.0287	13.0	2652	0.8958	0.8968	0.8963
0.0287	14.0	2856	1.0431	0.8854	0.8844
0.0244	15.0	3060	1.0537	0.8854	0.8844
0.0244	16.0	3264	0.8005	0.9054	0.9052
0.0244	17.0	3468	0.9819	0.8883	0.8893
0.02	18.0	3672	1.0702	0.8940	0.8928
0.02	19.0	3876	0.9675	0.8968	0.8957
0.0067	20.0	4080	0.9127	0.8968	0.8965
0.0067	21.0	4284	0.9818	0.9083	0.9075
0.0067	22.0	4488	0.9895	0.8940	0.8934
0.0074	23.0	4692	0.8589	0.9054	0.9054
0.0074	24.0	4896	1.0275	0.8997	0.8992
0.0139	25.0	5100	0.9546	0.9026	0.9021
0.0139	26.0	5304	0.9809	0.9083	0.9077
0.0074	27.0	5508	0.9914	0.9026	0.9020
0.0074	28.0	5712	0.9072	0.9054	0.9052
0.0074	29.0	5916	0.8984	0.9083	0.9081
0.0081	30.0	6120	0.9815	0.9083	0.9074
0.0081	31.0	6324	0.9143	0.8968	0.8969
0.003	32.0	6528	0.9652	0.9054	0.9044
0.003	33.0	6732	1.0522	0.9054	0.9045
0.003	34.0	6936	1.0332	0.9054	0.9045
0.0023	35.0	7140	1.0442	0.9054	0.9045

Framework versions

• Transformers 4.35.2
• Pytorch 1.13.1+cu117
• Datasets 2.14.5
• Tokenizers 0.15.2