Default embed-model isn't really set?

[Florents-Tselai]

Am I missing something in the following sequence or is this a bug where the default embed-model isn't really picked up, but works fine when it's provided explicitly?

$ llm embed-models default
sentence-transformers/all-MiniLM-L6-v2
$ llm embed -c "hello"
Error: You need to specify an embedding model (no default model is set)
$ llm embed -c "hello" -m sentence-transformers/all-MiniLM-L6-v2
[...OK...]

[simonw]

Thanks, this looks like a bug.

I replicated myself like this:

$ llm embed-models default
ada-002
$ llm embed -c "hello"
Error: You need to specify an embedding model (no default model is set)

[Florents-Tselai]

Yep, I just woke up & forked; I was about to check it out. Let me know if you'll be faster than me :) .

[simonw]

Fixed!