Just today, I finished a blog post (also my latest submission, felt like could be useful to some) about how to get something like this working in a bundle of something to run models, as well as a web UI for more easy interaction - in my case that was koboldcpp, which can run GGML, both on the CPU (with OpenBLAS) and on the GPU (with CLBlast). Thanks to Hugging Face, getting Metharme, WizardLM or other models is also…
Seems like serverless is the way to go for fast output while remaining inexpensive. e.g. https://replicate.com/stability-ai/stablelm-tuned-alpha-7b https://github.com/runpod/serverless-workers/tree/main/worke... https://modal.com/docs/guide/ex/falcon_gptq
GGML – AI at the Edge
121–130 of 246 posts
Re: GGML – AI at the Edge
#122Earlier quoted context omitted.
>> run LLaMA 30B at almost 3 tokens/s Please tell me your config! I have an i9-10900 with 32GB of ram that only gets .7 tokens/s on a 30B model
With a single NVIDIA 3090 and the fastest inference branch of GPTQ-for-LLAMA https://github.com/qwopqwop200/GPTQ-for-LLaMa/tree/fastest-i... , I get a healthy 10-15 tokens per second on the 30B models. IMO GGML is great (And I totally use it) but it's still not as fast as running the models on GPU for now.
I think it was originally designed to be easily embeddable—and most importantly, native code (i.e. not Python)—rather than competitive with GPUs.
I think it's just starting to get into GPU support now, but carefully.
Re: GGML – AI at the Edge
#123Earlier quoted context omitted.
> Please tell me your config! I have an i9-10900 with 32GB of ram that only gets .7 tokens/s on a 30B model Have you quantized it?
The model I have is q4_0 I think that's 4 bit quantized I'm running in Windows using koboldcpp, maybe it's faster in Linux?
That's correct, yeah. Q4_0 should be the smallest and fastest quantized model.
> I'm running in Windows using koboldcpp, maybe it's faster in Linux?
Possibly. You could try using WSL to test—I think both WSL1 and WSL2 are faster than Windows (but WSL1 should be faster than WSL2).
Re: GGML – AI at the Edge
#124Re: GGML – AI at the Edge
#125Just today, I finished a blog post (also my latest submission, felt like could be useful to some) about how to get something like this working in a bundle of something to run models, as well as a web UI for more easy interaction - in my case that was koboldcpp, which can run GGML, both on the CPU (with OpenBLAS) and on the GPU (with CLBlast). Thanks to Hugging Face, getting Metharme, WizardLM or other models is also…
Today with the new k-quants users are reporting that 30B models are working with 2-bit quantization on 16GB CPUs and GPUs [2]. That's enabling access to millions of consumers and the optimizations will only improve from there.
[1] https://old.reddit.com/r/LocalLLaMA/comments/13q6hu8/7b_perf...
[2] https://github.com/ggerganov/llama.cpp/pull/1684, https://old.reddit.com/r/LocalLLaMA/comments/141bdll/moneros...
Re: GGML – AI at the Edge
#126ggml and llama.cpp are such a good platform for local LLMs, having some financial backing to support development is brilliant. We should be concentrating as much as possible to do local inference (and training) based on privet data. I want a local ChatGPT fine tuned on my personal data running on my own device, not in the cloud. Ideally open source too, llama.cpp is looking like the best bet to achieve that!
Can LLaMA be used for commerical purposes though (might limit external contributors)? I believe, FOSS alternatives like DataBricks Dolly / Together RedPajama / Eluether GPT NeoX (et al) is where the most progress is likely to be at.
Re: GGML – AI at the Edge
#127I've always thought on the edge to be IoT type stuff. So running on embedded devices. But maybe that not the case?
My old job title had "edge" in it, and I still don't know what it's supposed to mean, although "not cloud" is a good approximation.
Re: GGML – AI at the Edge
#128I've always thought on the edge to be IoT type stuff. So running on embedded devices. But maybe that not the case?
Re: GGML – AI at the Edge
#129Earlier quoted context omitted.
The code is MIT licensed. If you don't agree with the direction the project is taking you can fork it and add whatever you want. I don't understand why this is so difficult for software developers with GitHub accounts to understand.
You've missed the point here more than I've seen anyone miss the point in a long time.
Re: GGML – AI at the Edge
#130ggml and llama.cpp are such a good platform for local LLMs, having some financial backing to support development is brilliant. We should be concentrating as much as possible to do local inference (and training) based on privet data. I want a local ChatGPT fine tuned on my personal data running on my own device, not in the cloud. Ideally open source too, llama.cpp is looking like the best bet to achieve that!
Can LLaMA be used for commerical purposes though (might limit external contributors)? I believe, FOSS alternatives like DataBricks Dolly / Together RedPajama / Eluether GPT NeoX (et al) is where the most progress is likely to be at.