Hugging Face's logo

adriansanz
/
2504v1

Text Classification
Transformers
Safetensors
roberta
Generated from Trainer
Inference Endpoints
Model card Files Community
2504v1 / README.md
adriansanz's picture
adriansanz
Update README.md
7f2b522 verified
preview code
|
raw
history blame contribute delete
6.91 kB
metadata
license: apache-2.0
base_model: projecte-aina/roberta-base-ca-v2-cased-te
tags:
  - generated_from_trainer
metrics:
  - accuracy
  - precision
  - recall
  - f1
model-index:
  - name: 2504v1
    results: []

2504v1

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: 0.5947
  • Accuracy: 0.8655
  • Precision: 0.8655
  • Recall: 0.8655
  • F1: 0.8655
  • Ratio: 0.5

Model description

Punto, change label 6 -------TRAIN------- Proporción de etiquetas en el conjunto de datos: Aigua: 36 muestras (5.88%) Consum, comerç i mercats: 36 muestras (5.88%) Cultura: 36 muestras (5.88%) Economia: 36 muestras (5.88%) Educació: 36 muestras (5.88%) Enllumenat públic: 36 muestras (5.88%) Esports: 36 muestras (5.88%) Habitatge: 36 muestras (5.88%) Horta: 36 muestras (5.88%) Medi ambient i jardins: 36 muestras (5.88%) Neteja de la via pública: 36 muestras (5.88%) Salut pública: 36 muestras (5.88%) Seguretat ciutadana i incivisme: 36 muestras (5.88%) Serveis socials: 36 muestras (5.88%) Tràmits: 36 muestras (5.88%) Urbanisme: 36 muestras (5.88%) Via pública i mobilitat: 36 muestras (5.88%)

-------VAL------- Proporción de etiquetas en el conjunto de datos: Aigua: 7 muestras (5.88%) Consum, comerç i mercats: 7 muestras (5.88%) Cultura: 7 muestras (5.88%) Economia: 7 muestras (5.88%) Educació: 7 muestras (5.88%) Enllumenat públic: 7 muestras (5.88%) Esports: 7 muestras (5.88%) Habitatge: 7 muestras (5.88%) Horta: 7 muestras (5.88%) Medi ambient i jardins: 7 muestras (5.88%) Neteja de la via pública: 7 muestras (5.88%) Salut pública: 7 muestras (5.88%) Seguretat ciutadana i incivisme: 7 muestras (5.88%) Serveis socials: 7 muestras (5.88%) Tràmits: 7 muestras (5.88%) Urbanisme: 7 muestras (5.88%) Via pública i mobilitat: 7 muestras (5.88%)

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: 8
  • seed: 42
  • 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: 10
  • label_smoothing_factor: 0.1

Training results

Training Loss Epoch Step Validation Loss Accuracy Precision Recall F1 Ratio
3.6209 0.2597 10 1.6277 0.5462 0.5476 0.5462 0.5430 0.4160
1.4156 0.5195 20 1.0896 0.5588 0.5620 0.5588 0.5531 0.6134
1.0016 0.7792 30 0.9251 0.5504 0.6083 0.5504 0.4811 0.8655
0.9148 1.0390 40 0.8180 0.6765 0.6912 0.6765 0.6701 0.3613
0.7958 1.2987 50 0.7074 0.7983 0.8038 0.7983 0.7974 0.5672
0.7218 1.5584 60 0.6919 0.8025 0.8216 0.8025 0.7995 0.6218
0.7019 1.8182 70 0.6693 0.8277 0.8383 0.8277 0.8264 0.4118
0.6805 2.0779 80 0.6229 0.8193 0.8232 0.8193 0.8188 0.5546
0.6206 2.3377 90 0.5833 0.8655 0.8665 0.8655 0.8655 0.4748
0.5979 2.5974 100 0.5642 0.8613 0.8614 0.8613 0.8613 0.5042
0.6115 2.8571 110 0.5634 0.8613 0.8614 0.8613 0.8613 0.5042
0.6016 3.1169 120 0.5447 0.8655 0.8665 0.8655 0.8655 0.5252
0.5514 3.3766 130 0.5601 0.8571 0.8588 0.8571 0.8570 0.5336
0.4678 3.6364 140 0.5717 0.8445 0.8475 0.8445 0.8442 0.5462
0.4962 3.8961 150 0.5684 0.8571 0.8575 0.8571 0.8571 0.5168
0.5214 4.1558 160 0.5573 0.8529 0.8536 0.8529 0.8529 0.5210
0.4962 4.4156 170 0.5686 0.8445 0.8475 0.8445 0.8442 0.5462
0.5032 4.6753 180 0.5525 0.8613 0.8616 0.8613 0.8613 0.4874
0.4593 4.9351 190 0.5747 0.8571 0.8581 0.8571 0.8571 0.5252
0.4335 5.1948 200 0.5919 0.8487 0.8488 0.8487 0.8487 0.5084
0.5023 5.4545 210 0.5854 0.8613 0.8626 0.8613 0.8612 0.4706
0.4399 5.7143 220 0.5728 0.8697 0.8719 0.8697 0.8696 0.5378
0.4182 5.9740 230 0.5737 0.8655 0.8665 0.8655 0.8655 0.5252
0.4337 6.2338 240 0.6013 0.8529 0.8536 0.8529 0.8529 0.5210
0.4046 6.4935 250 0.6200 0.8571 0.8575 0.8571 0.8571 0.5168
0.4304 6.7532 260 0.6106 0.8697 0.8698 0.8697 0.8697 0.5042
0.45 7.0130 270 0.6154 0.8655 0.8681 0.8655 0.8653 0.4580
0.3687 7.2727 280 0.6109 0.8655 0.8655 0.8655 0.8655 0.5
0.4102 7.5325 290 0.6118 0.8529 0.8536 0.8529 0.8529 0.5210
0.4197 7.7922 300 0.5969 0.8655 0.8656 0.8655 0.8655 0.4916
0.4874 8.0519 310 0.5794 0.8655 0.8656 0.8655 0.8655 0.4916
0.3694 8.3117 320 0.5777 0.8697 0.8704 0.8697 0.8697 0.5210
0.4029 8.5714 330 0.5828 0.8697 0.8700 0.8697 0.8697 0.5126
0.3946 8.8312 340 0.5860 0.8697 0.8698 0.8697 0.8697 0.5042
0.3991 9.0909 350 0.5864 0.8655 0.8655 0.8655 0.8655 0.5
0.3707 9.3506 360 0.5918 0.8697 0.8698 0.8697 0.8697 0.5042
0.3821 9.6104 370 0.5943 0.8655 0.8655 0.8655 0.8655 0.5
0.4135 9.8701 380 0.5947 0.8655 0.8655 0.8655 0.8655 0.5

Framework versions

  • Transformers 4.40.0
  • Pytorch 2.2.1+cu121
  • Datasets 2.19.0
  • Tokenizers 0.19.1