search

neuralmagic / deepsparse

Sparsity-aware deep learning inference runtime for CPUs
https://neuralmagic.com/deepsparse/
Other
2.94k stars 169 forks source link

Fix documentation about TextGeneration #1554

Closed luis5tb closed 5 months ago

luis5tb commented 5 months ago

TextGeneration takes model_path as parameter instead of model

mgoin commented 5 months ago

TextGeneration takes model_path as parameter instead of model

Hey @luis5tb model= should work for nightly and/or the upcoming 1.7, are you using the 1.6 release possibly?

luis5tb commented 5 months ago

TextGeneration takes model_path as parameter instead of model

Hey @luis5tb model= should work for nightly and/or the upcoming 1.7, are you using the 1.6 release possibly?

Yep using 1.6.1, following that documentation I hit it and saw that it was solved by adding model_path. Perhaps the code should take both then to keep backwards compatibility or add a note in the docs

luis5tb commented 5 months ago

Should I close this then?

mgoin commented 5 months ago

@luis5tb the nightly should both accept model_path and model, we just added model as a new alias that we'll be using going forward so backwards compatibility is good. It didn't make it into 1.6.1 though, so we'll be explicit once 1.7 comes out shortly