ddh0

Results 41 comments of ddh0

Hi @iwr-redmond, sorry for the delayed response. I didn't get notified about the issue for some reason. As of version 0.1.119, no, that parameter is not exposed. However I am...

Closed by latest commit. Thank you for your patience :)

This is a good idea. I'll come back to this thread when I start working on this specific functionality. I wonder if it would be possible to query the available...

Sorry, this feature is no longer planned

> Quantizing all LLAMA-2 models to 6- and 8-bits is lossless 😮‍💨

Hi @abetlen just checking in — would appreciate any input you have regarding this PR

Hey @abetlen . The original PR was opened in May. Are you still maintaining this repository?

> Discarding those backward compatibility changes, are there any other modifications that you still find worth adding in this PR? It allows execution of big-endian gguf files on big-endian hosts,...

> Could you keep only those changes? Done 👍