Ollama Installation and Configuration - Windows Systems

AI hands-on tutorials8mos agorelease AI Sharing Circle

38.3K 00

summary

In this section, learn how to do this on a Windows system Ollama The installation and configuration is divided into the following sections.

Visit the official website to complete the download directly
Environment variable configuration
Run Ollama
Verify that the installation was successful 🎉

I. Visit the official website to complete the download directly

Visit the official homepage
Ollama Download: https://ollama.com/download
Ollama official home page: https://ollama.com
Ollama official GitHub source code repository: https://github.com/ollama/ollama/
Wait for your browser to download the file OllamaSetup.exe, when it is finished, double-click the file, the following pop-up window will appear, click on the Install Just wait for the download to complete.
Once the installation is complete, you can see that Ollama is already running by default. You can find the Ollama logo in the bottom navigation bar and right click on it. Quit Ollama Exit Ollama or view logsThe

II. Configuration of environment variables

OllamaYou can install Ollama with one click on your computer like any other software, but it is recommended that you configure the system environment variables according to your actual needs. Below are the instructions for configuring environment variables for Ollama.

parameters	Labeling and Configuration
OLLAMA_MODELS	Indicates the directory where the model files are stored, the default directory isCurrent User Directoryassume (office)`C:\Users%username%.ollama\models`Windows system It is not recommended to put it on the C drive The disk can be placed on other disks (e.g. `E:\ollama\models`)
OLLAMA_HOST	Indicates the network address on which the ollama service listens, the default is127.0.0.1 If you want to allow other computers to access Ollama (e.g., other computers on a LAN), theRecommended settingsbe all right 0.0.0.0
OLLAMA_PORT	Indicates the default port that the ollama service listens on, which defaults to11434 If there is a port conflict, you can modify the settings to other ports (e.g.8080etc.)
OLLAMA_ORIGINS	Indicates the source of the request from the HTTP client, use half commas to separate the list, if the local use is not restricted, can be set to asterisks `*`
OLLAMA_KEEP_ALIVE	Indicates the survival time of a large model after it is loaded into memory, defaults to5mThat is, 5 minutes (e.g., a plain number 300 means 300 seconds, 0 means that the model is uninstalled as soon as the response to the request is processed, and any negative number means that the model is always alive) is recommended. 24h The model remains in memory for 24 hours, increasing access speeds.
OLLAMA_NUM_PARALLEL	Indicates the number of concurrent requests processed, defaults to1(i.e., single concurrent serial processing of requests) Recommendations adjusted to actual needs
OLLAMA_MAX_QUEUE	Indicates the request queue length, the default value is512It is recommended to adjust to the actual needs, requests exceeding the queue length will be discarded
OLLAMA_DEBUG	Indicates that the Debug log is output, which can be set to the following in the application development phase1(i.e., outputting detailed log information to facilitate troubleshooting)
OLLAMA_MAX_LOADED_MODELS	Indicates the maximum number of models loaded into memory at the same time, defaults to1(i.e. only 1 model can be in memory)

For beginners, we strongly recommend that you configure the OLLAMA_MODELS to change the model storage location. By default, Ollama models are stored in the user directory on the C drive under the .ollama/models folder, which takes up space on the C drive. Changing it to another partition will allow you to better manage your storage.

Step 1: Locate the entry point for setting system environment variables.

Method 1: Start->Settings->About->Advanced System Settings->System Properties->Environmental Variables.

Method 2: This computer -> Right click -> Properties -> Advanced System Settings -> Environment Variables.

Method 3: Start->Control Panel->System and Security->System->Advanced System Settings->System Properties->Environmental Variables.

Method 4: Win+R to open the Run window, enter sysdm.cpl, enter to open System Properties, select the Advanced tab, and click Environment Variables.

Step 2: Setup OLLAMA_MODELS Environment Variables (change the model storage location)

In " environment variable " Window's " System variables (S) " area (or the "User Variables (U)" area, depending on your needs), click the " new (W)... " button.
In " Variable name (N) " input box, type: OLLAMA_MODELS (Note the case, all capitalization is recommended).
In " Value of variable (V) " input box, enter the model storage path you want to set. For example, if you want to store the model on the E drive of the ollama\models folder, you can enter in the "Variable Value (V)": E:\ollama\models (Please modify the disk letter and folder path according to your actual situation).
- Attention: Make sure you enter the path Already Existing Folders or The parent directory of the folder you wish to create exists . Ollama may be automatically created when it is first run models folder, but it's a good idea to create the ollama folder to make sure the path is correct.
- Example: In order to store the model on the E-disk of the ollama\models folder, you can enter it in the "Variable Value (V)": E:\ollama\models
Click " recognize " button, close " New system variables " (or "New User Variable") window.

Attention: If you don't know how to set OLLAMA_MODELS environment variable, you can refer to the following example.

(Optional) Settings OLLAMA_HOST environment variable (modify the listening address).

If you need to make your Ollama service accessible to other devices on the LAN, you can configure the OLLAMA_HOST Environment variables.

Also in " environment variable " Window's " System variables (S) " area (or the "User Variables (U)" area), click the " new (W)... " button.
In " Variable name (N) " input box, type: OLLAMA_HOST
In " Value of variable (V) " input box, type 0.0.0.0:11434 (or the port number you want to specify; the default port is 11434).
- 0.0.0.0 Indicates listening to all network interfaces to allow LAN access.
- 11434 is the Ollama default port.
- Example: 0.0.0.0:11434
Click " recognize " button, close " New system variables " (or "New User Variable") window.

Step 3: Restart Ollama or PowerShell to make the environment variables take effect

Once the environment variables are set, you need to Restarting the Ollama Service or Restart your Command Prompt (CMD) or PowerShell window. in order for the new environment variable to take effect.

Restart the Ollama service. If you run the ollama serveprior Ctrl + C Stop and rerun ollama serveThe
Restart Command Prompt/PowerShell. Close all open windows and reopen new ones.

Step 4: Verify that the environment variables are in effect

reopen Command Prompt (CMD) or PowerShell The
validate (a theory) OLLAMA_MODELS: Type the following command and enter:

echo %OLLAMA_MODELS%

Output:E:\ollama\models (If you set the E:\ollama\models)

III. Running Ollama

Command line statements to start

ollama serve

The following error occurs when starting Ollama, because the Windows installation of Ollama will Default boot up (math.) genusOllama The service defaults to http://127.0.0.1:11434
Error: listen tcp 127.0.0.1:11434: bind: Only one usage of each socket address (protocol/network address/port) is normally permitted.

Solution:

(computer) shortcut key Win+X Open Task Manager and click 启动The following is an example of how to disable Ollama in thestepThe mission of Ollama ended in the middle of the year.
Re-use ollama serveOpen Ollama.

Verify successful startup:

(computer) shortcut key Win+RInput cmd, open a command line terminal.
importation netstat -aon|findstr 11434 View processes occupying port 11434.

netstat -aon|findstr 11434

Output: TCP 127.0.0.1:11434 0.0.0.0:0 LISTENING 17556
Shows that port 11434 is occupied by process 17556

A look at the process running shows that Ollama has started.

tasklist|findstr "17556"

The output is as follows:ollama.exe 17556 Console 1 31,856 K

IV. Verify successful installation🎉

Terminal Input:

ollama -h

The output is as follows: that means the installation was successful 🎉

Large language model runner
Usage:
ollama [flags]
ollama [command]
Available Commands:
serve       Start ollama
create      Create a model from a Modelfile
show        Show information for a model
run         Run a model
pull        Pull a model from a registry
push        Push a model to a registry
list        List models
ps          List running models
cp          Copy a model
rm          Remove a model
help        Help about any command
Flags:
-h, --help      help for ollama
-v, --version   Show version information