The Ministry of Digital Affairs is looking for volunteers to help prepare data for training the Ukrainian language model "Syaivo". Anyone can join. Knowledge of programming languages or technical experience is not required. Volunteers will review old archival documents and correct errors in texts recognized by the computer.
The Ministry of Digital Affairs is looking for volunteers to help prepare data for training the Ukrainian language model "Syaivo". Anyone can join. Knowledge of programming languages or technical experience is not required. Volunteers will review old archival documents and correct errors in texts recognized by the computer.
This was reported by the Ministry of Digital Affairs.
"Syavo" is being developed together with "Kyivstar". A large amount of materials have already been collected to train the model, but before using them to train AI, you need to make sure that the computer correctly recognized the texts in the scans.
You will need to work in the Label Studio service: compare the recognized text with the original scan, correct inaccuracies, and mark which fragments are printed and which are handwritten.
Those interested in joining need to fill out a short questionnaire. After that, coordinators will send instructions and provide access to the working platform.
Earlier, dev.ua wrote that the Ukrainian State Archives transferred 10 terabytes of data for training "Syayva". Among them are historical materials from different eras, state documents and scientific texts. In total, more than 50 partners have already joined the collection of data for the Ukrainian AI model, including universities, libraries and media.
