from freeGPT import Client
prompt = "Skomponuj mi dietę na najbliższy tydzień. Koniecznie odpisz po Polsku!"
resp = Client.create_completion("gpt4", prompt)
Before, the response message would contain unescaped unicode chars, like "\u0144". Now it's properly translated to "ń".
Similarly gpt3 and falcon (although the issue there is that requests decodes using latin1 instead of utf). Couldn't test alpaca - it always responds in English to me.
Quick test scenario:
Before, the response message would contain unescaped unicode chars, like "\u0144". Now it's properly translated to "ń".
Similarly gpt3 and falcon (although the issue there is that requests decodes using latin1 instead of utf). Couldn't test alpaca - it always responds in English to me.