Daniel Bevenius

@blaineam Could you give me some pointers to how to recreate these errors? I've renewed my apple developer account so I should be able to now. I though it would...

> Not sure if I have a strong opinion on this but removing hybrid sampling would reduce the complexity a bit I think (basically if we always set `--gpu-dist` we...

Would it be alright to limit the scope of this PR to only the following backend samplers: * logit bias * temperature * top_k * greedy * distribution And then...

> @danbev [7e98ebc](https://github.com/ggml-org/llama.cpp/pull/17004/commits/7e98ebcc6b2e0ceaa62350cf92c1d353ea61eff5) might have introduced a bug - I'm getting gibberish with backend sampling disabled. > > > Sorry about that, I'll look into it. It should be producing...

> Anything speaking against reserving appropriate buffers (and reusing n_output in case of `--backend_dist`) in `llama_context::output_reserve` when `--backend_sampling` and `--backend_dist` are specified? I've tried to address this in https://github.com/ggml-org/llama.cpp/pull/17004/commits/61ffe41dc1cdafb2c71b650c4265c22ec769f88b. I...

> And no matter what I do, the filename is always `Meta-Llama-3-8B-F16.gguf`. You can set the name of the output file using `--outfile` options: ```console ... options: -h, --help show...

@abc16361 I've taken a look into this and have a possible solution/suggestion. Full details/write-up can be found [here]( https://github.com/danbev/learning-ai/blob/main/notes/whisper-openblas-windows-issue.md#summary) What I believe is happening is that with the change to...

I'm happy to hear that this is working again, and sorry about the delay in providing a fix. Closing this issue now.