camembert-allocine / README.md
baptiste-pasquier's picture
add model
aee3262 verified
metadata
language:
  - fr
license: mit
tags:
  - generated_from_trainer
datasets:
  - allocine
widget:
  - text: Un film magnifique avec un duo d'acteurs excellent.
  - text: Grosse déception pour ce thriller qui peine à convaincre.
metrics:
  - accuracy
  - f1
  - precision
  - recall
model-index:
  - name: camembert-allocine
    results:
      - task:
          name: Text Classification
          type: text-classification
        dataset:
          name: allocine
          type: allocine
          config: allocine
          split: validation
          args: allocine
        metrics:
          - name: Accuracy
            type: accuracy
            value: 0.97535
          - name: F1
            type: f1
            value: 0.9749045558666326
          - name: Precision
            type: precision
            value: 0.9722814498933902
          - name: Recall
            type: recall
            value: 0.9775418538178848

camembert-allocine

This model is a fine-tuned version of camembert-base on the allocine dataset. It achieves the following results on the evaluation set:

  • Loss: 0.0928
  • Accuracy: 0.9754
  • F1: 0.9749
  • Precision: 0.9723
  • Recall: 0.9775

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: 5e-05
  • train_batch_size: 16
  • eval_batch_size: 16
  • seed: 42
  • gradient_accumulation_steps: 4
  • total_train_batch_size: 64
  • optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
  • lr_scheduler_type: linear
  • lr_scheduler_warmup_steps: 500
  • num_epochs: 3

Training results

Training Loss Epoch Step Validation Loss Accuracy F1 Precision Recall
0.1276 0.2 500 0.1187 0.9623 0.9622 0.9462 0.9787
0.1013 0.4 1000 0.0917 0.9683 0.9675 0.9725 0.9625
0.1254 0.6 1500 0.0889 0.9701 0.9698 0.9597 0.9801
0.1004 0.8 2000 0.0792 0.9716 0.9709 0.9727 0.9691
0.1149 1.0 2500 0.0762 0.9727 0.9723 0.9673 0.9773
0.0574 1.2 3000 0.0849 0.9733 0.9729 0.9679 0.9780
0.0394 1.4 3500 0.1026 0.9718 0.9715 0.9595 0.9839
0.0401 1.6 4000 0.1065 0.9698 0.9697 0.9528 0.9872
0.0458 1.8 4500 0.0834 0.9744 0.9739 0.9715 0.9764
0.0554 2.0 5000 0.0873 0.9719 0.9717 0.9594 0.9844
0.0516 2.2 5500 0.0928 0.9754 0.9749 0.9723 0.9775
0.0355 2.4 6000 0.1017 0.9744 0.9741 0.9642 0.9842
0.0227 2.6 6500 0.0983 0.9748 0.9743 0.9729 0.9757
0.0359 2.8 7000 0.0990 0.9747 0.9743 0.9665 0.9823
0.0384 3.0 7500 0.1001 0.9746 0.9742 0.9662 0.9824

Framework versions

  • Transformers 4.26.1
  • Pytorch 1.13.1+cu117
  • Datasets 2.10.1
  • Tokenizers 0.13.2