Obsidian with Ollama

in the previous video I introduced copilot plugin to you for running AI task for your notes in obsidian I mainly use cat GPT model and it was definitely making life easier however using cat GPT or other services for your notes means you're compromising your information on your notes in exchange of running AI task so if you are concerned about your data privacy today's video will be beneficial for you we're going to learn how to use Alama to to run local lmm on your machine and use it for obsidian so let's talk about a few terms first what is local lmm local lmm typically refers to running a large language model like llama 2 llama 3 mistr locally on your machine or server rather than accessing it through a cloud API provided by the company that create a model like open AI so what is the deal of using local lmm the first thing is you don't have to send your data to a third party cloud services when using the model this is critical as we are concerned about our notes being exposed and you don't have to pay subscription fees like when you use chat gbt Plus or Cloud a yet the performance of the model is directly limited to your computer's available resource however we can be a little smart to choose the the right model for your machine so what is ama ama is an open source project that streamlines the process of running running local lmms making it easy for everyone to use them the main advantage of using AMA is that it simplifies downloading managing and running models currently it supports llama 2 Lama 3 F3 Mistral jamama and it's free to run this models on your machine so let's dive in into how to use this for our obsidian so let's install AMA first go to the website download Based on your operating system once it's downloaded you should be able to run the AMA commands as a headup AMA require you to know a little bit about terminal but I will cover all your need so don't get scared so when you're ready open up the terminal first thing you can do is to type Lama this will show you the available commands out of all we need only least pull run and serve for today's setup I will explain each of them right now the first command is pull AMA pull model name this will try to download model if it doesn't exist on your machine yet once you download it you can use AMA list to check out the download models on your machine and you can also check out models page in the official website to see what are the available models they provide here models and they have llama 3 V3 Mist Gemma COA Etc so back to the terminal the next command I want to let you know is AMA run Lama R modal name like this well downloads and start a chat with that model right inside the terminal this is amazing as you can have conversation with AI directly in the terminal so hello word in Python for example but that is not what we want for obsidian we want a local server for running AI inside OB cidian to do so we need to use Serve command this can be a little tricky as you have to specify origion to avoid any course issue simply type Lama origin app obsidian MD llama serve then we're done on the server side I'll put this command down in the description so you can easily copy and paste it without any headaches next thing is go to obsidian copilot open up the obsidian and change the default model from the previous one to AMA local and scroll all the way down to AMA model and specify model you want to run with for my case I will I will use F3 before you close this setting make sure to save and reload your configuration so that obsidian will reflect your changes now you're good to go there is one note if your computer has smaller Ram I'd recommend to try smaller model otherwise you will get freezed and frustrated with the speed and want to compromise your data for the speed that cloud provides for my case when I use llama 3 literally my computer stops working but I was able to run F3 smoothly with Mag Air 8 gig M1 setup you can use obsidian with Alama like you use with CET gbt T summarization explain like I'm 5-year-old translate make templates easier watch my another video to take advantage of AI for your notes yet without concerns of your notes being exposed to anyone my primary reason for using obsidian was Data privacy so I finally fell at home with Lama we covered how to install Lama and use stem with obsidian for your notes I think a lot of services currently ask your data in exchange for their services however your notes idea thoughts may be the most important value so protect them while still utilizing what AI offers for your productivity if you like this video please make sure to subscribe this channel see you in another video

Channel: The Writer Dev 글쓰는 개발자

Views: 12,046

Keywords:

Id: 9YYB8a_ehc4

Length: 6min 32sec (392 seconds)

Published: Tue Apr 30 2024