GGchen1997

Hi everyone, thank you for the fantastic work. I have a question regarding the sampling process. Since flow matching aims to learn the vector field of an ODE, the sampling...

Scale of importance weight

1

Hi, thanks for your interesting and valuable work. When I run the code, I get the importance weight scale varies: from 5.4e-52 to 6.5e40. Is this normal?

num-start and num-end

3

Thanks for the fantastic work you present! https://huggingface.co/GSAI-ML/LLaDA-8B-Instruct/blob/main/special_tokens_map.json In the tokenizer, we have num-start and num-end, two special tokens. I want to ask whether the two tokens have been used...