tairov / llama2.mojo

Inference Llama 2 in one file of pure 🔥
https://www.modular.com/blog/community-spotlight-how-i-built-llama2-by-aydyn-tairov
MIT License
2.09k stars 139 forks source link

✂️ remove call to external os library #54

Closed shroominic closed 10 months ago

shroominic commented 10 months ago

I had this issue #53 when trying to run the demo in the README. Removing the need for the external python os call fixed it and I think this is generally an improvement to the project! Have not measured it but this will probaply decrease the size binary when building and it removes the need of having a working python installation.

Great project tho now I can experiment with it!

shroominic commented 10 months ago
Screenshot 2023-10-20 at 22 46 29

This is my result now - 86t/s not bad!