Eric Hartford
Eric Hartford
Maybe you could try using nightly cuda and pytorch?
Oh yeah I had the same thing
Ok but, is it better to support hugging face instead of having to copy the dataset to s3? Aws charges for ingress and egress
Fair; but beyond my skill level. I will ask my network and see if I can find any insight.
we know that it's much more efficient training with Scatter MoE and we would like to benefit from the cost savings
this is feature request, not a bug that could be reproduced. The academic paper I am requesting is linked above.
Very interested in this
I also need CPU-only quantization support I have lots of RAM, and I don't care if it is slow. Surely GPU is not a requirement, mathematically speaking.
Yeah I know that it's possible to change it through some menus options. I wanted it to be more directly in the chat screen and easy to quickly change it....
 if you could add a flag there, that shows the system message box System message, it shouldn't be global. It's per-conversation, not global. also I found it using that...