I'm having a hard time trying to update a demo that I have based on your mlc-llm container. I moved to JP6 and the server I wrote to use a quantized version of Vicuna seems to be broken for some lib incompatibility.
how did you get those wheels? did you build them yourself? I can't install newer versions from here: https://mlc.ai/wheels
as it seems they are not compatible (of course, the cuda ones are built for x86!!)
Hello @dusty-nv !
I'm having a hard time trying to update a demo that I have based on your mlc-llm container. I moved to JP6 and the server I wrote to use a quantized version of Vicuna seems to be broken for some lib incompatibility.
I opened a bug to the mlc-ai team, but I'm a bit stuck at the moment: https://github.com/mlc-ai/mlc-llm/issues/2512
how did you get those wheels? did you build them yourself? I can't install newer versions from here: https://mlc.ai/wheels as it seems they are not compatible (of course, the cuda ones are built for x86!!)
Do you have any clue about what's going on?
Thank you so much in advance!