Eric Curtin comments

Results 479 comments of


                                            Eric Curtin

guided_decoding is not defined when using GRPOTrainer

@ochafik kindly made the initial change to common_chat_templates_init in llama-run. This change ensures we don't pass nullptr to the function.

guided_decoding is not defined when using GRPOTrainer

I merged this anyway to get rid of the seg fault people were reporting.

relabel-etc: Split remount into, remount and relabel-etc

Incomplete and untested right now

relabel-etc: Split remount into, remount and relabel-etc

I like the idea of using setfiles or something similar, less code to maintain... I think selinux=0, etc. is easy to handle in systemd .service file: ConditionKernelCommandLine=!selinux=0 We may be...

Invalid args when installing with pipx

Hi @derrickqi2003 can you test if this change fixes your issue: https://github.com/containers/ramalama/pull/306 we haven't thoroughly tested docker support yet. Just podman support and without containers at all. It should only...

Implement whisper.cpp

We added a basic version of whisper.cpp to the Container image here: https://github.com/containers/ramalama/pull/49

Implement whisper.cpp

There's still more work for both `--runtime whisper.cpp` and `--runtime vllm` ... It can be tracked under this issue or somewhere else.

Implement whisper.cpp

I would say just this to start: `ramalama --runtime whisper.cpp run ggml-large-v3-turbo.bin jfk.wav` which would just perform a fairly standard whisper.cpp command. No interactive support, it's not the same as...

Implement whisper.cpp

@p5 kindly set up renovate for us.

Implement whisper.cpp

I meant something like this, corrected: `ramalama --runtime whisper.cpp run ggml-large-v3-turbo.bin jfk.wav`