basilepp19/cpv-it5

model_5M

This model is a fine-tuned version of gsarti/it5-large on a dataset of Common Procurement Vocabulary (CPV) codes.

Model description

The model is trained on 3.2M pairs of Italian tender descriptions and the corresponding CPV code.

Here an example:

{"source": "lavori lavori di pavimentazione delle vie san martino e santa Maddalena", "target": "45262321-7 - lavori di pavimentazione"}

Intended uses & limitations

This model can generate a CPV code given an Italian tender description.

Training and evaluation data

Training data are taken form the ANAC website.

Training procedure

Training hyperparameters

The following hyperparameters were used during training:

• learning_rate: 5e-05
• train_batch_size: 16
• eval_batch_size: 8
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 3.0

Training results

Framework versions

• Transformers 4.26.1
• Pytorch 1.13.1+cu117
• Datasets 2.9.0
• Tokenizers 0.13.2