Metrics in Chinese?
#4
by
exoplanet
- opened
Neither the blog post nor the model page has any metrics in Chinese. It's okay if the model didn't do well in Chinese, but how bad really, we'd need to know to make a call on whether or not to use this model for Chinese use-cases. Honestly, I expect more transparency from a reputable lab like Liquid AI.
Regarding the chinese metrics, we have extended the nanobeir benchmark for other languages, and we had people available to review the translation in Korean and Japanese. By then, we didn't have anyone to review the translations to Chinese. That was the main reason
fernandofernandes
changed discussion status to
closed