New ask Hacker News story: Ask HN: How to get started with local language models?
Ask HN: How to get started with local language models?
2 by sandwichukulele | 0 comments on Hacker News.
I remember using Talk to a Transformer in 2019 and making little Markov chains for silly text generation. I've lurked in the g /lmg/ threads, installed the Umbrel LlamaGPT on a Raspberry Pi, and ran webGPU WebGPT (GPT-2) locally. But I don't know how anything works or how to do anything besides following installation instructions on GitHub. HuggingFace still confuses me; there's so much stuff out there to read through, and I have been lost since the release of the llama models. I heard about Mistral. I tried using the Mozilla Llamafiles since it was meant to be easy for the Mixtral-8x7B-Instruct, but apparently, windows won't run an EXE bigger than 4GB, so I have to do something about the weights, and I freeze here. I know where to look and how to learn about any other technology, but I'm completely lost for how to learn about local models when everything is moving so fast. I'm missing something fundamental. How can I understand these technologies?
2 by sandwichukulele | 0 comments on Hacker News.
I remember using Talk to a Transformer in 2019 and making little Markov chains for silly text generation. I've lurked in the g /lmg/ threads, installed the Umbrel LlamaGPT on a Raspberry Pi, and ran webGPU WebGPT (GPT-2) locally. But I don't know how anything works or how to do anything besides following installation instructions on GitHub. HuggingFace still confuses me; there's so much stuff out there to read through, and I have been lost since the release of the llama models. I heard about Mistral. I tried using the Mozilla Llamafiles since it was meant to be easy for the Mixtral-8x7B-Instruct, but apparently, windows won't run an EXE bigger than 4GB, so I have to do something about the weights, and I freeze here. I know where to look and how to learn about any other technology, but I'm completely lost for how to learn about local models when everything is moving so fast. I'm missing something fundamental. How can I understand these technologies?
Comments
Post a Comment