? ollama Part 1. Running it from a local terminal: Linux(WSL 2), macOS & ollama version update
- ollama Part 2. Running it from a local browser : open-webui
- ollama Part 3. Running it online (via my own domain) in a browser
- ollama Part 4. Applying retrieval-augmented generation (RAG)
- ollama Part 5. Adding image recognition
- (in preparation) ollama Part 6. Applying the MOE (Mixture of Experts) approach
ollama local terminal on linux(wsl 2) in windows 11
1. Installing Linux (Ubuntu) via the Windows Subsystem (WSL 2)
1) From Windows search, run [Turn Windows features on or off]
2) Configure the options
3) Reboot.
4) From Windows search, search for [cmd] and run [Terminal] in Administrator mode.
5) In the terminal, run wsl --install -d ubuntu.
-> You'll set an ID and password partway through.
6) In the terminal, wsl --update
7) In the terminal, wsl --set-default-version 2
8) In the terminal, wsl --set-version Ubuntu 2
9) Run Ubuntu
.png)
2. Installing ollama on the local machine (Ubuntu)
1) Go to the Ollama site.
Download Ollama on Linux
Download Ollama on Linux
ollama.com
2) In the Ubuntu console (terminal or CLI), run curl -fsSL https://ollama.com/install.sh | sh
3) In the Ubuntu console, run ollama serve.
-> A key gets issued.
4) In the Ubuntu console, run ollama run mistral *If it isn't installed, this installs it; if it is installed, this runs it.
-> Run or install the model
*Running ollama run [model_name] runs the model if it's there; if not, it installs it.
* The page that introduces models and explains how to install them: https://ollama.com/library
library
Get up and running with large language models, locally.
ollama.com
5) Other commands
- In the Ubuntu console: /? :help
- In the Ubuntu console: /show :list of main commands
- In the Ubuntu console: /show info :example list of main commands
- In the Ubuntu console: /bye
6) Example run
-> The question, in Korean, was "Please tell me about the capital of South Korea."
.png)
-> DeepL translation
Also called "the city of the calm morning" or simply Seoul Metropolitan City, Seoul is the capital and largest metropolis of the Republic of Korea. It is located in the northwestern part of the country and forms the heart of the Seoul Capital Area, which includes the surrounding Incheon Metropolitan City and Gyeonggi Province.
Seoul has a long history of more than 2,000 years, starting as a small trading town. Throughout its history, it has been an important cultural, political, and economic center of Korea. Today Seoul is a leading global city and a major hub for education, research, finance, and technology.
Seoul has many iconic landmarks that draw millions of tourists every year. Must-visit attractions include:
1. Gyeongbokgung Palace: the largest and most famous of the five Joseon-era royal palaces, built between 1392 and 1897, showcasing Korean traditional architecture, gardens, and culture.
2. Namsan: a popular getaway with hiking trails, a cable car, and various scenic spots, including the famous love locks at Namsan Love Lock Plaza. 3. Myeong-dong: a shopping district famous for street food, cosmetics shops, and traditional Korean medicine markets.
3. Insa-dong: an area famous for antique shops, tea houses, and galleries, where you can get a glimpse of traditional Korean culture.
4. Dongdaemun Design Plaza: a modern architectural landmark with various museums, artworks, and shopping facilities…
ollama local terminal on mac
1. Installing ollama on the local machine (Mac)
1) Go to ollama.
2) Download.
3) Search for [Terminal] and run it.
4) Run ollama
ollama
5) Install (= run) a model*If it isn't installed, this installs it; if it is installed, this runs it.
ollama run gemma
#Reference: an introduction to Google Gemma (link to Google's official blog)
구글의 최첨단 오픈 모델 ‘젬마(Gemma)’를 공개합니다
젬마는 제미나이 모델을 개발하기 위해 활용됐던 것과 동일한 연구와 기술을 바탕으로 책임감 있는 AI 개발을 위해 탄생하였습니다.
blog.google
6) Search the models you've installed so far
ollama list
ollama version update
About a week ago? While installing gemma on Windows (WSL2) I got an error message, and looking it up, it was a bug introduced by an ollama version update. https://github.com/ollama/ollama/issues/2646
Defect: EOF on running with Gemma:7b · Issue #2646 · ollama/ollama
OS: Mac M1 Pro $ ollama run gemma:7b pulling manifest pulling 2c5f288be750... 100% ▕████████████████████████████████████████████▏ 4.8 GB pulling 097a36493f71... 100
github.com
So I tried to install the new version manually.. but I couldn't seem to find clear instructions even after searching. I did find them eventually.. but in case there are others wandering around in a similar state, I'm leaving this in a post. The method is... apparently you just install it again (overwriting it) lol https://github.com/ollama/ollama/issues/763
update local ollama install · Issue #763 · ollama/ollama
How to update ollama cli locally with latest features?
github.com
For reference, on Mac, as of yesterday.. when I ran it again after a long time, it asked me whether I wanted to update the version, so I was able to update easily. By now perhaps a feature for automatically checking and updating the version has also made it into the Windows and Linux builds..
