LlamaEdge / LlamaEdge

The easiest & fastest way to run customized and fine-tuned LLMs locally or on the edge

Home Page:https://llamaedge.com/

Geek Repo:Geek Repo

Github PK Tool:Github PK Tool

Feature Request: Is it possible to see the output realtime?

alexcardo opened this issue · comments

Summary

We used to see how the model answering the question especially in the UI mode. Please add this feature.

Thank you!

Appendix

No response

Hi @alexcardo

Do you mean supporting streaming output in the web chatbot U? If so, we're working on it. Stay tuned.

Yes, this is exactly what I mean. I'm waiting for this!

Also, it would be great to tune up settings visually. I mean temperature, top_k, top_p, etc.

I've got this that I can use it from console while starting the server, but it also would be great to see it in the UI.

Thank you.

By the way, as I opened the issue. In the CLI version, there is no multiline input. Correct me if I'm wrong.

I mean, for instance, I want to act like this:

Summarize this:

bla bla bla....

In LLAMA CPP, there is a --multiline-input, where I can use a "\" symbol.

By the way, as I opened the issue. In the CLI version, there is no multiline input. Correct me if I'm wrong.

I mean, for instance, I want to act like this:

Summarize this:

bla bla bla....

In LLAMA CPP, there is a --multiline-input, where I can use a "\" symbol.

PR #91 provides support for multiline inputs in the console.

Can you give an example of how to use the multiline input? Once I type Enter, the bot starts answering the question.

I used this instructionhttps://www.secondstate.io/articles/run-llm-sh/to install LlamaEdge

I did curl -LO https://github.com/second-state/LlamaEdge/releases/latest/download/llama-chat.wasm hoping that there are these changes:

https://github.com/second-state/LlamaEdge/pull/91/commits/bb78cf8f9d44db9b187871b327e3be3468d65270#diff-46cfefd927062cfabcabb14bc2a3cf0b5dab89b0470471c7a39a98521270447f

Either I don't realize how this implementation work or there should be some instruction. I would be happy to see it.

@alexcardo At the end of each line, just use "", then press Enter. You'll get a new line. Hope the video can help you.

multiline-inputs.mov

Looks like this issue is resolved. Is it fine to close it?