Hugging Face's logo

bigscience
/
bloom

Text Generation Transformers PyTorch TensorBoard Safetensors 46 languages
bloom Eval Results Carbon Emissions Inference Endpoints text-generation-inference 4 papers
Model card Files Metrics Community
275

Why was no slavic language included in the training dataset?

#56
by brabecjan91 - opened
Discussion
brabecjan91

It would seem natural to me to include at least a few languages from the Slavic family given that there was apparently a process to make the languages in the training dataset diverse.

brabecjan91 changed discussion title from Why was no slavic language included in training dataset? to Why was no slavic language included in the training dataset?
SaulLu
 BigScience Workshop org

A good pointer that could answer your question may be @yjernite tweet answering "why wasn't language X included in the @BigScienceLLM training data" 🤗

TimeRobber
 BigScience Workshop org

Closing as this seems to have been resolved. Thank you!

TimeRobber changed discussion status to closed

Sign up or log in to comment