Tae Hwan Jung

저도 학부생으로써 비슷한 질문을 드리고 싶습니다! :D 저에게 100% 공감할수 있는 글을 써주신 @litcoderr @SPark9625 님께 감사드립니다.

Thanks for sharp caught Could you contribute for that? Also it contains jupyter file corresponding to code

**[TODO]** https://www.reddit.com/r/MachineLearning/comments/d0u73h/p_the_basic_distribution_probability_tutorial_for/ezedide?utm_source=share&utm_medium=web2x Add conjugate prior relationships after Review 1. Beta is the conjugate prior of binomial and Bernoulli (p). 2. Dirichlet for multinomial and categorical (probability vector). 3. Gaussian of...

@armoreal Which language do you want? Is it English?

@armoreal First, Existing gpt-2 models are only supported in English. https://github.com/openai/gpt-2/issues/31 If you want to train your language, I recommend you to read original gpt, gpt-2 paper. Please See [Improving...

@armoreal I think this repository can be trainable https://github.com/openai/finetune-transformer-lm but, There are no dataset related to your langauge and computer resource I think.. In gpt-2 paper, they explained what is...

@armoreal See code and paper more detail  1. Text-Predict in here : https://github.com/openai/finetune-transformer-lm/blob/master/train.py#L176, 3.1 Unsupervisedpre-training 2. Task classification in here : https://github.com/openai/finetune-transformer-lm/blob/master/train.py#L193, 3.2 Supervisedfine-tuning `L3(C) = L2(C) + λ∗L1(C)`...

Overall, There is code related with training. so you can train. BUT Dataset and Computer power maybe problem :( **Please do not close this issue for everyone!**

Yes I agree with you Thanks

> * np.float32 Thanks for your reply First, I cannot create a schema dimension of float type. `TileDBError: [TileDB::ArraySchema] Error: Cannot set domain; Dense arrays do not support dimension datatype...