This model was trained from scratch using the Fairseq toolkit on a combination of Spanish-Catalan datasets, up to 92 million sentences. Additionally, the model is evaluated on several public datasecomprising 5 different domains (general, adminstrative, technology, biomedical, and news).
Intended uses and limitations
You can use this model for machine translation from Spanish to Catalan.
The model was trained on a combination of the following datasets:
Dataset
Sentences
Tokens
DOCG v2
8.472.786
188.929.206
El Periodico
6.483.106
145.591.906
EuroParl
1.876.669
49.212.670
WikiMatrix
1.421.077
34.902.039
Wikimedia
335.955
8.682.025
QED
71.867
1.079.705
TED2020 v1
52.177
836.882
CCMatrix v1
56.103.820
1.064.182.320
MultiCCAligned v1
2.433.418
48.294.144
ParaCrawl
15.327.808
334.199.408
Total
92.578.683
1.875.910.305
Training procedure
Data preparation
All datasets are concatenated and filtered using the mBERT Gencata parallel filter and cleaned using the clean-corpus-n.pl script from moses, allowing sentences between 5 and 150 words.
Before training, the punctuation is normalized using a modified version of the join-single-file.py script from SoftCatalà
Tokenization
All data is tokenized using sentencepiece, with 50 thousand token sentencepiece model learned from the combination of all filtered training data. This model is included.
Hyperparameters
The model is based on the Transformer-XLarge proposed by Subramanian et al.
The following hyperparamenters were set on the Fairseq toolkit:
Hyperparameter
Value
Architecture
transformer_vaswani_wmt_en_de_bi
Embedding size
1024
Feedforward size
4096
Number of heads
16
Encoder layers
24
Decoder layers
6
Normalize before attention
True
--share-decoder-input-output-embed
True
--share-all-embeddings
True
Effective batch size
96.000
Optimizer
adam
Adam betas
(0.9, 0.980)
Clip norm
0.0
Learning rate
1e-3
Lr. schedurer
inverse sqrt
Warmup updates
4000
Dropout
0.1
Label smoothing
0.1
The model was trained using shards of 10 million sentences, for a total of 8.000 updates. Weights were saved every 1000 updates and reported results are the average of the last 6 checkpoints.
This work was funded by the Spanish State Secretariat for Digitalization and Artificial Intelligence (SEDIA)
Disclaimer
Click to expand
The models published in this repository are intended for a generalist purpose and are available to third parties. These models may have bias and/or any other undesirable distortions.
When third parties, deploy or provide systems and/or services to other parties using any of these models (or using systems based on these models) or become users of the models, they should note that it is their responsibility to mitigate the risks arising from their use and, in any event, to comply with applicable regulations, including regulations regarding the use of Artificial Intelligence.
In no event shall the owner of the models (SEDIA – State Secretariat for Digitalization and Artificial Intelligence) nor the creator (BSC – Barcelona Supercomputing Center) be liable for any results arising from the use made by third parties of these models.
Los modelos publicados en este repositorio tienen una finalidad generalista y están a disposición de terceros. Estos modelos pueden tener sesgos y/u otro tipo de distorsiones indeseables.
Cuando terceros desplieguen o proporcionen sistemas y/o servicios a otras partes usando alguno de estos modelos (o utilizando sistemas basados en estos modelos) o se conviertan en usuarios de los modelos, deben tener en cuenta que es su responsabilidad mitigar los riesgos derivados de su uso y, en todo caso, cumplir con la normativa aplicable, incluyendo la normativa en materia de uso de inteligencia artificial.
En ningún caso el propietario de los modelos (SEDIA – Secretaría de Estado de Digitalización e Inteligencia Artificial) ni el creador (BSC – Barcelona Supercomputing Center) serán responsables de los resultados derivados del uso que hagan terceros de estos modelos.
