lewtun's picture
lewtun HF staff
Add evaluation results on the default config of xsum
07e9e1b
|
raw
history blame
1.74 kB
metadata
tags:
  - summarization
language:
  - it
metrics:
  - rouge
model-index:
  - name: summarization_ilpost
    results:
      - task:
          type: summarization
          name: Summarization
        dataset:
          name: xsum
          type: xsum
          config: default
          split: test
        metrics:
          - name: ROUGE-1
            type: rouge
            value: 12.1495
            verified: true
          - name: ROUGE-2
            type: rouge
            value: 1.6326
            verified: true
          - name: ROUGE-L
            type: rouge
            value: 10.493
            verified: true
          - name: ROUGE-LSUM
            type: rouge
            value: 10.5515
            verified: true
          - name: loss
            type: loss
            value: 2.5110762119293213
            verified: true
          - name: gen_len
            type: gen_len
            value: 18.9924
            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