bert-text-classification-car-evaluation

This model is a fine-tuned version of distilbert/distilbert-base-uncased on the Car Evaluation Dataset. You can always find it here in Hugging Face Hub.

It achieves the following results on the evaluation set:

• Loss: 0.0090
• F1: 1.0
• Roc Auc: 1.0
• Accuracy: 1.0

Model description

The model is designed for text classification tasks on the Car Evaluation Dataset. It is a fine-tuned version of the DistilBERT model, aiming to predict car evaluation categories based on textual information.

Intended uses & limitations

Intended Uses

• Car evaluation category prediction based on textual information.
• Text classification tasks related to the car evaluation domain.

Limitations

• The model's performance is specifically tuned for the Car Evaluation Dataset; its generalization to other tasks or datasets might be
• limited.
• It may not perform optimally on text from different domains or with substantially different linguistic characteristics.

Training and evaluation data

The model was trained on the Car Evaluation Dataset, which includes textual descriptions of cars along with corresponding evaluation categories.

Training procedure

Training hyperparameters

The following hyperparameters were used during training:

• learning_rate: 2e-05
• train_batch_size: 32
• eval_batch_size: 16
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 50
• early_stopping: 10

Training results

The model stopped training at the 29th epoch, achieving the following results on the evaluation set:

Training Loss	Epoch	Step	Validation Loss	F1	Roc Auc	Accuracy
0.469	1.0	33	0.3870	0.6812	0.7874	0.6812
0.3686	2.0	66	0.3724	0.6812	0.7874	0.6812
0.3455	3.0	99	0.3243	0.6921	0.7787	0.6058
0.2809	4.0	132	0.2348	0.8148	0.8720	0.7971
0.1939	5.0	165	0.1762	0.8571	0.9034	0.8522
0.1609	6.0	198	0.1655	0.8734	0.9145	0.8696
0.1395	7.0	231	0.1302	0.9163	0.9406	0.9043
0.1261	8.0	264	0.1133	0.9161	0.9396	0.9014
0.097	9.0	297	0.1180	0.8986	0.9324	0.8754
0.0906	10.0	330	0.1212	0.9052	0.9391	0.8870
0.0851	11.0	363	0.0947	0.9078	0.9357	0.8899
0.0792	12.0	396	0.0933	0.9320	0.9551	0.9188
0.073	13.0	429	0.0783	0.9277	0.9527	0.9217
0.0586	14.0	462	0.0737	0.9577	0.9696	0.9420
0.0682	15.0	495	0.0855	0.9312	0.9512	0.9188
0.0625	16.0	528	0.0869	0.9391	0.9594	0.9246
0.0567	17.0	561	0.0653	0.9525	0.9705	0.9420
0.0513	18.0	594	0.0576	0.9666	0.9773	0.9565
0.0463	19.0	627	0.0655	0.9595	0.9739	0.9449
0.047	20.0	660	0.0485	0.9608	0.9734	0.9478
0.0379	21.0	693	0.0406	0.9825	0.9855	0.9739
0.0338	22.0	726	0.0274	0.9827	0.9894	0.9739
0.0325	23.0	759	0.0215	0.9942	0.9952	0.9913
0.0254	24.0	792	0.0251	0.9913	0.9932	0.9884
0.0266	25.0	825	0.0212	0.9884	0.9923	0.9826
0.0203	26.0	858	0.0170	0.9913	0.9932	0.9884
0.0193	27.0	891	0.0149	0.9986	0.9995	0.9971
0.0204	28.0	924	0.0140	0.9971	0.9971	0.9942
0.0162	29.0	957	0.0094	1.0	1.0	1.0
0.0157	30.0	990	0.0103	1.0	1.0	1.0
0.0139	31.0	1023	0.0084	1.0	1.0	1.0
0.0125	32.0	1056	0.0076	1.0	1.0	1.0
0.0105	33.0	1089	0.0067	1.0	1.0	1.0
0.0091	34.0	1122	0.0058	1.0	1.0	1.0
0.009	35.0	1155	0.0064	1.0	1.0	1.0
0.0081	36.0	1188	0.0053	1.0	1.0	1.0
0.0074	37.0	1221	0.0050	1.0	1.0	1.0
0.008	38.0	1254	0.0050	1.0	1.0	1.0
0.0077	39.0	1287	0.0053	1.0	1.0	1.0

Framework versions

• Transformers 4.38.2
• Pytorch 2.2.1
• Datasets 2.18.0
• Tokenizers 0.15.2

Acknowledgments

This model is built upon the distilbert/distilbert-base-uncased pre-trained model and utilizes the Hugging Face Transformers library. Special thanks to the creators of the Car Evaluation Dataset for providing the training and evaluation data.

Contact Information

For any questions or inquiries, please contact the model developer:

Name: Henriquee

Hugging Face: Henriquee

License

This model is released under the MIT License. See the LICENSE file for more details.