IzwiIzwi

Manual Download: Gemma 3 1B Instruct

Gemma 3 1B Instruct (Gemma-3-1b-it) is a powerful compact language model from Google that requires manual download due to licensing requirements.

Why Manual Download?

Google's Gemma models require you to:

  1. Have a Hugging Face account
  2. Accept the Gemma license agreement
  3. Use authenticated downloads

This cannot be done automatically by Izwi.

Prerequisites

Create a Hugging Face Account

If you don't have one, sign up at huggingface.co.

Accept the Gemma License

  1. Visit the Gemma 3 1B model page
  2. Click Agree and access repository
  3. Read and accept the license terms

Note: License approval is usually instant but may take a few minutes.

Step 1: Install Hugging Face CLI

Choose one method:

Option 1: Using pipx (Recommended)

pipx install huggingface_hub

Option 2: Using pip

python3 -m pip install --upgrade "huggingface_hub[cli]"

Step 2: Authenticate

Log in to Hugging Face from your terminal:

huggingface-cli login

When prompted:

  1. Go to huggingface.co/settings/tokens
  2. Create a new token with Read permissions
  3. Paste the token into your terminal

Verify you're logged in:

huggingface-cli whoami

Step 3: Download the Model

macOS

huggingface-cli download google/gemma-3-1b-it \ --repo-type model \ --local-dir "$HOME/Library/Application Support/izwi/models/Gemma-3-1b-it"

Linux

huggingface-cli download google/gemma-3-1b-it \ --repo-type model \ --local-dir "$HOME/.local/share/izwi/models/Gemma-3-1b-it"

Windows (PowerShell)

huggingface-cli download google/gemma-3-1b-it ` --repo-type model ` --local-dir "$env:APPDATA\izwi\models\Gemma-3-1b-it"

The download is approximately 2.5 GB and may take several minutes depending on your connection.

Step 4: Verify Installation

Restart Izwi or refresh the model list:

izwi list --local

You should see Gemma-3-1b-it in the list.

Load the model:

izwi models load Gemma-3-1b-it

Troubleshooting

"Access to model is restricted"

You haven't accepted the license yet:

  1. Visit huggingface.co/google/gemma-3-1b-it
  2. Click Agree and access repository
  3. Wait a few minutes for approval
  4. Try the download again

"401 Client Error: Unauthorized"

Your token doesn't have the right permissions:

  1. Go to huggingface.co/settings/tokens
  2. Create a new token with Read access
  3. Run huggingface-cli login again with the new token

Model not appearing in Izwi

  1. Check the model was downloaded to the correct directory
  2. Verify the folder name is exactly Gemma-3-1b-it
  3. Restart the Izwi server

Download interrupted

The Hugging Face CLI automatically resumes. Just run the same download command again.

Using Gemma 3 1B

Once loaded, you can use Gemma for chat:

izwi chat --model Gemma-3-1b-it

Or via the web UI at http://localhost:8080/chat.

See Also