camembert-bank-post-classifier

It classifies posts on an internal social media in 7 categories :

• Annonce RH & Vie de l'Entreprise
• Communication de la Direction & Stratégie
• Formation & Développement
• Information & Alerte Pratique
• Projet, Produit & Innovation
• RSE & Engagement Citoyen
• Événement Externe & Partenariat

Model description

This model is a fine-tuned version of camembert-base on an unknown dataset.

Intended uses & limitations

Classifying posts under 7 categories.

Training and evaluation data

It achieves the following results on the evaluation set:

• Loss: 0.5783
• Accuracy: 0.8682
• F1: 0.8663

Training procedure

Training hyperparameters

The following hyperparameters were used during training:

• learning_rate: 2e-05
• train_batch_size: 16
• eval_batch_size: 16
• seed: 42
• optimizer: Use OptimizerNames.ADAMW_TORCH_FUSED with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
• lr_scheduler_type: linear
• num_epochs: 8

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy	F1
1.557	1.0	124	1.2381	0.7181	0.6649
1.0415	2.0	248	0.8830	0.8316	0.8254
0.6258	3.0	372	0.6725	0.8560	0.8523
0.4723	4.0	496	0.6103	0.8540	0.8516
0.3092	5.0	620	0.5921	0.8600	0.8583
0.2244	6.0	744	0.5722	0.8641	0.8624
0.1482	7.0	868	0.5719	0.8641	0.8626
0.1493	8.0	992	0.5783	0.8682	0.8663

Framework versions

• Transformers 4.56.1
• Pytorch 2.8.0+cu126
• Datasets 4.0.0
• Tokenizers 0.22.0