search

ngxson / wllama

WebAssembly binding for llama.cpp - Enabling in-browser LLM inference
https://ngxson.github.io/wllama/examples/basic/
MIT License
231 stars 5 forks source link

Would it be possible to release a new version? #67

Closed flatsiedatsie closed 3 weeks ago

flatsiedatsie commented 3 weeks ago

The "use the first chunk name only" feature would be very useful.

ngxson commented 3 weeks ago

Sorry I was quite busy recently. In fact, I wanted to see if I can fix the problem when loading files > 2GB in this release, but seems like it's more complicated than I thought, so let's delay it until next version.

The current main branch looks good to me now. I'll try to release it in next hours.

flatsiedatsie commented 3 weeks ago

Thanks!

I was quite busy recently

Good!

the problem when loading files > 2GB

Which problem is that?

ngxson commented 3 weeks ago

It's not possible to load a single big file (need to split into small files) ==> https://github.com/ngxson/wllama/pull/52#issuecomment-2138242650

Ability to load big file can be useful when we want to pick a gguf file from local

ngxson commented 3 weeks ago

v1.11.0 released on npm

Changelog: https://github.com/ngxson/wllama/releases/tag/1.11.0

flatsiedatsie commented 3 weeks ago

Thank you