Gavin S

Looking for something already posted to docker hub to avoid that build step on systems without build infrastructure in place.

I don't remember the context for this comment/request 😅. Maybe some of the thumbs and rockets do?

https://kaitai.io/ https://ide.kaitai.io/ ```yaml meta: id: compobj endian: le encoding: ASCII seq: - id: unk_skip size: 28 - id: av_size type: u4 - id: app_version type: str size: av_size - id:...

There's also this craziness: https://www.youtube.com/watch?v=GLjKQHMU57k And this, although calling this a twist might be a stretch. https://www.rbracing-rsr.com/turbo/wiring_pix/service_loops.jpg Both of these may very well be their own mechanical thing unnecessary to...

Not updated very often, and not down to county level. If we're just going to point to other resources to get a complete dataset, why not contribute to the other...

The LLMs in langchain have a token count function. Easier to use that than bringing in another library that doesn't know the specifics of the model. That said I noticed...

> `from_toktoken_encoder` and `from_huggingface_tokenizer` classmethods both use a token-based length_function My particular ask is to have it called as a class method so that each LLM can provide it's appropriate...

Alright, yeah I don't know why that didn't push through before. This whole thing should be squashed anyway so it should be one nice clean commit when done.

Easy enough to detect if it's passed as a value, or a list. Shouldn't break backwards compatibility.

The package listed at this URL appears to contain a patch that resolves the vulnerability. https://packages.ubuntu.com/source/bionic/beep Going straight to the debian package that fixed it did not have a patch,...