Fairseq
Italian
Catalan
fdelucaf commited on
Commit
2c1826c
1 Parent(s): b26c9cd

Update README.md

Browse files
Files changed (1) hide show
  1. README.md +16 -1
README.md CHANGED
@@ -109,30 +109,45 @@ The following hyperparameters were set on the Fairseq toolkit:
109
  | Warmup updates | 8000 |
110
  | Dropout | 0.1 |
111
  | Label smoothing | 0.1 |
 
112
  The model was trained for a total of 19.000 updates. Weights were saved every 1000 updates and reported results are the average of the last 4 checkpoints.
 
113
  ## Evaluation
 
114
  ### Variable and metrics
115
- We use the BLEU score for evaluation on the Flores test set: [Flores-101](https://github.com/facebookresearch/flores),
 
 
116
  ### Evaluation results
 
117
  Below are the evaluation results on the machine translation from Catalan to Italian compared to [Softcatalà](https://www.softcatala.org/) and [Google Translate](https://translate.google.es/?hl=es):
 
118
  | Test set | SoftCatalà | Google Translate |mt-aina-it-ca|
119
  |----------------------|------------|------------------|---------------|
120
  | Flores 101 dev | 25,4 | **30,4** | 27,5 |
121
  | Flores 101 devtest |26,6 | **31,2** | 27,7 |
122
  | NTREX | 29,3 | **33,5** | 30,7 |
123
  | Average | 27,1 | **31,7** | 28,6 |
 
124
  ## Additional information
 
125
  ### Author
126
  Language Technologies Unit (LangTech) at the Barcelona Supercomputing Center.
 
127
  ### Contact information
128
  For further information, send an email to <langtech@bsc.es>
 
129
  ### Copyright
130
  Copyright Language Technologies Unit at Barcelona Supercomputing Center (2023)
 
131
  ### Licensing information
132
  This work is licensed under a [Apache License, Version 2.0](https://www.apache.org/licenses/LICENSE-2.0)
 
133
  ### Funding
134
  This work was funded by the Departament de la Vicepresidència i de Polítiques Digitals i Territori de la Generalitat de Catalunya within the framework of Projecte AINA.
 
135
  ### Disclaimer
 
136
  <details>
137
  <summary>Click to expand</summary>
138
  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.
 
109
  | Warmup updates | 8000 |
110
  | Dropout | 0.1 |
111
  | Label smoothing | 0.1 |
112
+
113
  The model was trained for a total of 19.000 updates. Weights were saved every 1000 updates and reported results are the average of the last 4 checkpoints.
114
+
115
  ## Evaluation
116
+
117
  ### Variable and metrics
118
+
119
+ We use the BLEU score for evaluation on the [Flores-101](https://github.com/facebookresearch/flores), and [NTREX](https://github.com/MicrosoftTranslator/NTREX) evaluation datasets.
120
+
121
  ### Evaluation results
122
+
123
  Below are the evaluation results on the machine translation from Catalan to Italian compared to [Softcatalà](https://www.softcatala.org/) and [Google Translate](https://translate.google.es/?hl=es):
124
+
125
  | Test set | SoftCatalà | Google Translate |mt-aina-it-ca|
126
  |----------------------|------------|------------------|---------------|
127
  | Flores 101 dev | 25,4 | **30,4** | 27,5 |
128
  | Flores 101 devtest |26,6 | **31,2** | 27,7 |
129
  | NTREX | 29,3 | **33,5** | 30,7 |
130
  | Average | 27,1 | **31,7** | 28,6 |
131
+
132
  ## Additional information
133
+
134
  ### Author
135
  Language Technologies Unit (LangTech) at the Barcelona Supercomputing Center.
136
+
137
  ### Contact information
138
  For further information, send an email to <langtech@bsc.es>
139
+
140
  ### Copyright
141
  Copyright Language Technologies Unit at Barcelona Supercomputing Center (2023)
142
+
143
  ### Licensing information
144
  This work is licensed under a [Apache License, Version 2.0](https://www.apache.org/licenses/LICENSE-2.0)
145
+
146
  ### Funding
147
  This work was funded by the Departament de la Vicepresidència i de Polítiques Digitals i Territori de la Generalitat de Catalunya within the framework of Projecte AINA.
148
+
149
  ### Disclaimer
150
+
151
  <details>
152
  <summary>Click to expand</summary>
153
  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.