adriansanz/080524_15ep_02

080524_epoch_5

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: 0.5972
• Accuracy: 0.8445
• Precision: 0.8448
• Recall: 0.8445
• F1: 0.8445
• Ratio: 0.4874

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: 10
• eval_batch_size: 2
• seed: 47
• gradient_accumulation_steps: 2
• total_train_batch_size: 20
• 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
0.4518	0.1626	10	0.6633	0.8361	0.8469	0.8361	0.8348	0.4118
0.4418	0.3252	20	0.6798	0.8277	0.8279	0.8277	0.8277	0.5126
0.5709	0.4878	30	0.7447	0.8193	0.8367	0.8193	0.8170	0.3866
0.6645	0.6504	40	0.6229	0.8487	0.8487	0.8487	0.8487	0.5
0.6606	0.8130	50	0.6014	0.8445	0.8446	0.8445	0.8445	0.5042
0.5763	0.9756	60	0.5972	0.8445	0.8448	0.8445	0.8445	0.4874

                                     precision      recall    f1-score  top1-score  top2-score  top3-score good1-score good2-score     support

0 Aigua 0.632 0.545 0.585 0.545 0.818 0.955 0.955 0.955 22 1 Consum, comerç i mercats 0.103 0.571 0.174 0.571 0.714 0.857 0.714 0.714 7 2 Cultura 0.500 0.750 0.600 0.750 0.750 0.750 0.750 0.750 8 3 Economia 0.211 0.500 0.296 0.500 0.875 1.000 0.875 0.875 8 4 Educació 0.438 0.636 0.519 0.636 0.818 1.000 1.000 1.000 11 5 Enllumenat públic 0.833 0.851 0.842 0.851 0.936 0.979 0.979 0.979 47 6 Esports 0.562 0.750 0.643 0.750 0.917 1.000 1.000 1.000 12 7 Habitatge 0.208 0.385 0.270 0.385 0.615 0.923 0.692 0.846 13 8 Horta 0.000 0.000 0.000 0.000 0.444 0.556 0.556 0.556 9 9 Medi ambient i jardins 0.429 0.559 0.485 0.559 0.729 0.915 0.915 0.915 59 10 Neteja de la via pública 0.686 0.238 0.353 0.238 0.505 0.772 0.762 0.762 101 11 Salut pública 0.135 0.292 0.184 0.292 0.708 0.792 0.708 0.708 24 12 Seguretat ciutadana i incivisme 0.727 0.471 0.571 0.471 0.588 0.765 0.706 0.706 34 13 Serveis socials 0.333 0.667 0.444 0.667 0.889 0.889 0.889 0.889 9 14 Tràmits 0.395 0.395 0.395 0.395 0.884 0.907 0.907 0.907 43 15 Urbanisme 0.379 0.172 0.237 0.172 0.453 0.641 0.578 0.578 64 16 Via pública i mobilitat 0.778 0.778 0.778 0.778 0.846 0.889 0.864 0.867 279

                       macro avg       0.432       0.504       0.434       0.504       0.735       0.858       0.815       0.824         750
                    weighted avg       0.610       0.557       0.559       0.557       0.739       0.853       0.825       0.829         750
                        accuracy       0.557
                      error rate       0.443

Framework versions

• Transformers 4.40.1
• Pytorch 2.2.1+cu121
• Datasets 2.19.1
• Tokenizers 0.19.1