Hi, Mentor! I read all the discussion related to this exercise and took the advice to restart my kernel and rerun the cells. But I still got the same result ---- a bunch of “eu”. I passed the unite tests. I wonder where the mistake is from. Any advices? Many thanks!
Sincerely,
Shining
Hi @goldfish Seem like you have pass the test.
the bunch of eu showing you variation between temperature in the translation model. There is no issue as per what you have shared.
Do you understand the significance of temperature here?
Temperature parameter determines the creativity and diversity of the text generated by the model.
A higher temperature value (e.g., 1.5) leads to more diverse and creative text, while a lower value (e.g., 0.5) results in more focused and deterministic text. So when temperature value of 0 was used, results varied to be too restrictive leading to bunches of eu’s
Regards
DP
Thank you for your reply! I remember you explained the meaning of temperature in other threads about the same topic. I remember some one got a smaller size of translation tokens. So I want to know why I had 50 "eu"s. Shouldn’t I get 5 or 6 "eu"s? I am lost.
Your question has a bit answer to your doubt? i suppose you already when through the comment you commented on the other thread. Did you notice the variation in logits?
model is being trained on randomness of weight which will vary from your model to my model, resulting in different output.
Try re-run your model and then translation do you get the same bunch of EU’s 
1 Like