trishullab / LibraryAugmentedSymbolicRegression.jl / 13914428920

34%

push

github

Auto-download LLM (#43)

* adding llamafile exec capability

* format :/

* update deps

* update async call

* update serve_llm

* add: llamafile support.
llamafile will now start up with correct env variiable.

* Updating logging verbosity.
This helps address parsing issues that show up when using smaller LLMs.

* Update example

* formatting :/

* add flag passing option

* remove default port due to type instability

* formatting and resolve type instability.

* remove output type signature

* add LLM_FLAGS only if positive length

* for MacOS, we need to invoke the server with bashotherwise exec doesn't understand how to call the llamafile.

* update prompts

46 of 61 new or added lines in 3 files covered. (75.41%)

417 of 914 relevant lines covered (45.62%)

31536390.91 hits per line