xiaoyun wu

@fwcd I am also interested in incremental compilation, and what is the newest pointers on this? Thanks.

Is this fixed already?

@zinuga same here

I have the same question here.

> When you add extra tokens, the embedding dimensions of some layers change. You have to add those layers in modules_to_save. Depending on your model, the modules_to_save can change. Please...

instead of support more models, I think we should get this working first? I am interested in this. There is a S-Lora out there already.

> Yes, we are currently reading papers about Retrieval Augmented LM to find out what training/adaptation setup to RAG is better suited for TinyLlama. It we be great if you...

https://github.com/yaodongC/awesome-instruction-dataset @jzhang38 Just in case you did not see this.

instruction tuning is not zero-shot (prompt engineering can be).

Presentation: Chatbot from first principle Presenter: Sean Wu, Framely Abstraction: What is chatbot anyways, who pays for it, and what is implied on the first principle from these axiom like...