maddes8cht
maddes8cht
I think `Ctrl+C` is a bad choice to interrupt the output, because `Ctrl+C` also stops the program. It has happened to me that I pressed `Ctrl+C` at the wrong moment...
### Feature request The models that can be downloaded with the chat client are ggml files, which can also be loaded with ggml's llama.cpp main program. ### Motivation I know...
After the CUDA refactor PR #1703 by @JohannesGaessler was merged i wanted to try it out this morning and measure the performance difference on my ardware. I use my standard...
The Roadmap needs to be updated again. #15661 updated the roadmap for 2023 near the end of the year. It' s 2024 now. Isn't a radmap supposed to show the...
The parsing of the parameters is broken. Basically, parameters were only parsed correctly to arguments if it was the last argument of the command line. For most parameter errors in...
I have always been irritated (also in Llama.cpp) by the fact that the help text cannot be piped. Neither a `falcon-main --help |less` (the help is meanwhile 85 lines long)...
In the past few days, the server-example from llama.cpp has become a really useful piece of software - so much so that for many things it could replace the main...
The readme sugests: > Windows: you can find up-to-date binaries in the [Actions](https://github.com/magiblot/turbo/actions?query=branch:master+event:push) page. But the artifacts on the actions page have expired. Currently there are no binaries to download....
### Please confirm these before moving forward. - [x] I have searched for my feature proposal and have not found a work-in-progress/duplicate/resolved/discarded issue. - [X] This proposal is a completely...