Xuan-Rui Fan

原因在[這裡](https://github.com/Morizeyao/GPT2-Chinese/issues/88#issuecomment-824588058), 如果無法解決可以嘗試在生成文本時把[UNK] token的logits設為`-float('Inf')`

I found out why. My `len(logits_lists)` was 2, and my `logits_percentage` was 0.25. In this line: https://github.com/timoschick/pet/blob/21d32de975a911bfa0261827c9bd23dc4f0e4aa2/pet/modeling.py#L714 my `num_logits_lists`, after applying `round()`, become 0. Afterwards, my `logits_lists` after this line:...

try get_substances('NSC231854', 'name')? It is a substance record on pubchem website.

> When will Ollama support be available? I'm trying to do it with [forked](https://github.com/microdev1/guidance/tree/ollama) Ollama, but I'm getting {G|Number|G} if it's not being handled correctly. The Ollama initialization seems to...

From what I know, for a model to work with guidance, it needs to provide guidance role start and role end tags, e.g., `\n` and `\n` for Phi3 Small and...

> Just thinking out loud a bit... But could structured state (in some form -- not necessarily the rough implementation in my open PR) allow us to call `tokenizer.apply_chat_template` directly,...

I just wrote the scripts [here](https://gist.github.com/xruifan/d4041b8ebd7f6965f35d373964dcd12b) not sure if this is what you wanted 🙏