adriansanz/te-zsc-authentic

AUTH_300524_epoch_4

This model is a fine-tuned version of projecte-aina/roberta-base-ca-v2-cased-te on the None dataset. It achieves the following results on the evaluation set:

• Loss: 0.4656
• Accuracy: 0.9038
• Precision: 0.9047
• Recall: 0.9038
• F1: 0.9038
• Ratio: 0.4760

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: 16
• eval_batch_size: 16
• seed: 47
• gradient_accumulation_steps: 2
• total_train_batch_size: 32
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• lr_scheduler_warmup_ratio: 0.06
• lr_scheduler_warmup_steps: 4
• num_epochs: 1
• label_smoothing_factor: 0.1

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy	Precision	Recall	F1	Ratio
0.4294	0.0354	10	0.5003	0.9018	0.9020	0.9018	0.9018	0.5100
0.386	0.0708	20	0.5308	0.8938	0.8952	0.8938	0.8937	0.4699
0.4424	0.1062	30	0.4881	0.8998	0.9000	0.8998	0.8998	0.4900
0.42	0.1416	40	0.4916	0.9068	0.9091	0.9068	0.9067	0.4629
0.418	0.1770	50	0.4905	0.8968	0.8968	0.8968	0.8968	0.4950
0.4402	0.2124	60	0.5034	0.8988	0.9027	0.8988	0.8986	0.4509
0.4141	0.2478	70	0.5085	0.9028	0.9061	0.9028	0.9026	0.4549
0.4836	0.2832	80	0.4875	0.9028	0.9029	0.9028	0.9028	0.4910
0.4361	0.3186	90	0.4876	0.8998	0.8998	0.8998	0.8998	0.4980
0.45	0.3540	100	0.4985	0.8938	0.8938	0.8938	0.8938	0.5040
0.4648	0.3894	110	0.5236	0.8858	0.8954	0.8858	0.8851	0.4218
0.4714	0.4248	120	0.5009	0.8888	0.8888	0.8888	0.8888	0.5010
0.4628	0.4602	130	0.4971	0.8868	0.8871	0.8868	0.8867	0.4850
0.4513	0.4956	140	0.4971	0.8968	0.9003	0.8968	0.8966	0.4529
0.4905	0.5310	150	0.4873	0.8938	0.8969	0.8938	0.8936	0.4559
0.4875	0.5664	160	0.4760	0.8948	0.8948	0.8948	0.8948	0.4950
0.4593	0.6018	170	0.4818	0.8918	0.8918	0.8918	0.8918	0.4960
0.403	0.6372	180	0.4927	0.8928	0.8936	0.8928	0.8927	0.4770
0.4838	0.6726	190	0.5039	0.8958	0.9001	0.8958	0.8955	0.4479
0.4512	0.7080	200	0.4913	0.8978	0.9009	0.8978	0.8976	0.4559
0.4415	0.7434	210	0.4874	0.8988	0.8989	0.8988	0.8988	0.4930
0.5317	0.7788	220	0.4786	0.9018	0.9021	0.9018	0.9018	0.4860
0.4718	0.8142	230	0.4746	0.9008	0.9041	0.9008	0.9006	0.4549
0.473	0.8496	240	0.4686	0.9028	0.9044	0.9028	0.9027	0.4689
0.499	0.8850	250	0.4689	0.9028	0.9031	0.9028	0.9028	0.4870
0.5655	0.9204	260	0.4661	0.9068	0.9074	0.9068	0.9068	0.4810
0.4583	0.9558	270	0.4654	0.9048	0.9057	0.9048	0.9048	0.4770
0.4734	0.9912	280	0.4656	0.9038	0.9047	0.9038	0.9038	0.4760

Framework versions

• Transformers 4.41.1
• Pytorch 2.3.0+cu121
• Datasets 2.19.1
• Tokenizers 0.19.1