metadata

tags:
  - summarization
language:
  - it
metrics:
  - rouge
model-index:
  - name: summarization_ilpost
    results:
      - task:
          type: summarization
          name: Summarization
        dataset:
          name: ARTeLab/ilpost
          type: ARTeLab/ilpost
          config: ARTeLab--ilpost
          split: test
        metrics:
          - name: ROUGE-1
            type: rouge
            value: 28.4008
            verified: true
          - name: ROUGE-2
            type: rouge
            value: 13.9951
            verified: true
          - name: ROUGE-L
            type: rouge
            value: 24.1571
            verified: true
          - name: ROUGE-LSUM
            type: rouge
            value: 26.0996
            verified: true
          - name: loss
            type: loss
            value: 1.6566967964172363
            verified: true
          - name: gen_len
            type: gen_len
            value: 18.9439
            verified: true
datasets:
  - ARTeLab/ilpost

summarization_ilpost

This model is a fine-tuned version of gsarti/it5-base on IlPost dataset for Abstractive Summarization.

It achieves the following results:

Loss: 1.6020
Rouge1: 33.7802
Rouge2: 16.2953
Rougel: 27.4797
Rougelsum: 30.2273
Gen Len: 45.3175

Usage

from transformers import T5Tokenizer, T5ForConditionalGeneration
tokenizer = T5Tokenizer.from_pretrained("ARTeLab/it5-summarization-ilpost")
model = T5ForConditionalGeneration.from_pretrained("ARTeLab/it5-summarization-ilpost")

Training hyperparameters

The following hyperparameters were used during training:

learning_rate: 5e-05
train_batch_size: 6
eval_batch_size: 6
seed: 42
optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
lr_scheduler_type: linear
num_epochs: 4.0

Framework versions

Transformers 4.12.0.dev0
Pytorch 1.9.1+cu102
Datasets 1.12.1
Tokenizers 0.10.3