Howto ollama: Difference between revisions

From Vidalinux Wiki
Jump to navigation Jump to search
Created page with "= install packages = * archlinux install nvidia drivers: install ollama dependencies: pacman -S ollama ollama-cuda nvidia-container-toolkit docker docker-compose"
 
No edit summary
Tag: Manual revert
 
(20 intermediate revisions by the same user not shown)
Line 2: Line 2:
* archlinux
* archlinux
install nvidia drivers:
install nvidia drivers:
   
  https://wiki.vidalinux.org/index.php?title=Howto_install_nvidia_drivers#archlinux
install ollama dependencies:
install ollama dependencies:
  pacman -S ollama ollama-cuda nvidia-container-toolkit docker docker-compose
  pacman -S ollama ollama-cuda nvidia-container-toolkit docker docker-compose
= service =
copy systemd ollama daemon to /etc/systemd/system/:
cp /usr/lib/systemd/system/ollama.service /etc/systemd/system/
edit /etc/systemd/system/ollama.service and add the following:
Environment="OLLAMA_HOST=0.0.0.0"
reload systemd daemon:
systemctl daemon-reload
restart ollama.service
systemctl restart ollama.service
verify ollama is listening on all interfaces:
netstat -tnulp|grep 11434
= ollama =
to search for supported models on ollama:
https://ollama.com/search
to pull a model:
ollama pull deepseek-r1:8b
to list models installed:
ollama list
to run a model:
ollama run deepseek-r1:8b
to remove a model:
ollama rm deepseek-r1:8b
to show info about model:
ollama show deepseek-r1:8b
= create your own ssl cert =
create rsa key:
openssl genrsa -out server.key 3072
create certificate csr:
openssl req -new -key server.key -out server.csr
fill the following blanks:
Country Name (2 letter code) []: '''US'''
State or Province Name (full name) []: '''Puerto Rico'''
Locality Name (eg, city) []: '''San Juan'''
Organization Name (eg, company) []: '''Vidalinux.com Corp.'''
Organizational Unit Name (eg, section) []: '''Linux Consulting'''
Common Name (eg, your name or your server's hostname) []: '''chat.example.com'''
Email Address []: '''chat@example.com'''
Please enter the following 'extra' attributes
to be sent with your certificate request
A challenge password []: '''just press enter'''
An optional company name []: '''just press enter'''
create the certificate:
openssl x509 -req -days 365 -in server.csr -signkey server.key -out server.crt
= openwebui =
clone git repo:
git clone https://github.com/vidalinux/ollama.git
copy server.key and server.crt to nginx/ssl/
cd ollama/openwebui
mkdir nginx/ssl
cp server.key server.crt nginx/ssl/
deploy services using docker-compose:
docker-compose up -d
add your domain to /etc/hosts file:
echo "127.0.0.1 chat.example.com"
to enter openwebui interface:
https://chat.example.com
= vscodium =
for development install vscodium, fork of vscode without telemetry:
pacman -S vscodium --noconfirm
when installed, go to extensions and install:
continue
select your own ai model:
ollama

Latest revision as of 02:12, 21 August 2025

install packages

  • archlinux

install nvidia drivers:

https://wiki.vidalinux.org/index.php?title=Howto_install_nvidia_drivers#archlinux

install ollama dependencies:

pacman -S ollama ollama-cuda nvidia-container-toolkit docker docker-compose

service

copy systemd ollama daemon to /etc/systemd/system/:

cp /usr/lib/systemd/system/ollama.service /etc/systemd/system/

edit /etc/systemd/system/ollama.service and add the following:

Environment="OLLAMA_HOST=0.0.0.0"

reload systemd daemon:

systemctl daemon-reload

restart ollama.service

systemctl restart ollama.service

verify ollama is listening on all interfaces:

netstat -tnulp|grep 11434

ollama

to search for supported models on ollama:

https://ollama.com/search

to pull a model:

ollama pull deepseek-r1:8b

to list models installed:

ollama list

to run a model:

ollama run deepseek-r1:8b

to remove a model:

ollama rm deepseek-r1:8b

to show info about model:

ollama show deepseek-r1:8b

create your own ssl cert

create rsa key:

openssl genrsa -out server.key 3072

create certificate csr:

openssl req -new -key server.key -out server.csr

fill the following blanks:

Country Name (2 letter code) []: US
State or Province Name (full name) []: Puerto Rico
Locality Name (eg, city) []: San Juan
Organization Name (eg, company) []: Vidalinux.com Corp.
Organizational Unit Name (eg, section) []: Linux Consulting
Common Name (eg, your name or your server's hostname) []: chat.example.com
Email Address []: chat@example.com
Please enter the following 'extra' attributes
to be sent with your certificate request
A challenge password []: just press enter
An optional company name []: just press enter

create the certificate:

openssl x509 -req -days 365 -in server.csr -signkey server.key -out server.crt

openwebui

clone git repo:

git clone https://github.com/vidalinux/ollama.git

copy server.key and server.crt to nginx/ssl/

cd ollama/openwebui
mkdir nginx/ssl
cp server.key server.crt nginx/ssl/

deploy services using docker-compose:

docker-compose up -d

add your domain to /etc/hosts file:

echo "127.0.0.1 chat.example.com"

to enter openwebui interface:

https://chat.example.com

vscodium

for development install vscodium, fork of vscode without telemetry:

pacman -S vscodium --noconfirm

when installed, go to extensions and install:

continue

select your own ai model:

ollama