# How to Fine-tune LLMs in VS Code with Unsloth & Colab GPUs

You can now fine-tune LLMs directly from Visual Studio Code (VSCode), locally or by using Google Colab's extension. In this guide, you’ll learn how to use the open-source training [repo: Unsloth](https://github.com/unslothai/unsloth), to connect any [fine-tuning notebook](https://unsloth.ai/docs/get-started/unsloth-notebooks) in VS Code to a Colab runtime, so you can train on your local or free Colab GPU. You can also view our video tutorial [here](#video-tutorial).

{% stepper %}
{% step %}

### VS Code and Colab Tutorial:

To begin we will need to have:

* Installed [VS Code](https://code.visualstudio.com/). Git (for cloning the notebook repo) should be installed by default.
* A **Google account** (to authenticate with Colab)
* Recommended: **Jupyter** extension (most VS Code setups already have it)
  {% endstep %}

{% step %}

#### Install the Colab extension in VS Code

1. Open **Extensions** in VS Code (`Ctrl+Shift+X` / `Cmd+Shift+X`)
2. Search for **“Colab”** and install the **Google Colab** extension

<figure><img src="https://3215535692-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FxhOjnexMCB3dmuQFQ2Zq%2Fuploads%2FJ1f5BS4EK0ok5QLy6vz5%2Fcolab_img_1.png?alt=media&#x26;token=1faa7aac-c016-4c31-90ba-ccad655244e1" alt="" width="375"><figcaption></figcaption></figure>
{% endstep %}

{% step %}

#### Open an Unsloth notebook

1. Clone the Unsloth [notebooks repository](https://github.com/unslothai/notebooks):

```bash
git clone https://github.com/unslothai/notebooks
cd notebooks/nb
```

<figure><img src="https://3215535692-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FxhOjnexMCB3dmuQFQ2Zq%2Fuploads%2FwRL8pvgtLRpz5rcmS6vy%2FScreenshot%202026-02-18%20at%206.29.16%E2%80%AFPM.png?alt=media&#x26;token=7a72c2ec-6f94-4015-8355-27e76ba9b974" alt="" width="375"><figcaption></figcaption></figure>

2. Open your desired notebook. Unsloth supports most models including [embedding](https://unsloth.ai/docs/basics/embedding-finetuning), [TTS](https://unsloth.ai/docs/basics/text-to-speech-tts-fine-tuning). For example, we'll use Qwen3-4B RL: `nb/Qwen3_(4B)-GRPO.ipynb`

<figure><img src="https://3215535692-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FxhOjnexMCB3dmuQFQ2Zq%2Fuploads%2FleZtIzyflSgi1o1QcZDU%2Fcolab_img_2.png?alt=media&#x26;token=805ebe8b-cb50-49a4-83d4-0daf96ad21c1" alt="" width="375"><figcaption></figcaption></figure>
{% endstep %}

{% step %}

#### Select a kernel and choose Colab

In the notebook toolbar, click **Select Kernel**, then choose **Colab**

<figure><img src="https://3215535692-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FxhOjnexMCB3dmuQFQ2Zq%2Fuploads%2FleGtAKSVehrlspPGvDy2%2Fcolab_img_3.png?alt=media&#x26;token=ba3d8f1d-9cbb-44ed-8e52-8d0e52280607" alt="" width="375"><figcaption></figcaption></figure>
{% endstep %}

{% step %}

#### Add a new Colab server

After choosing **Colab**, you’ll see a dropdown with server options.

1. Click **+ Add New Colab Server**
2. The first time, a browser window may open for Google authentication
   * Log in, grant access, then return to VS Code

<figure><img src="https://3215535692-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FxhOjnexMCB3dmuQFQ2Zq%2Fuploads%2FZcDsB2i0aUStIKEeFTvF%2Fcolab_img_4.png?alt=media&#x26;token=a7c3642f-db35-4215-933f-99c17ebc92a2" alt="" width="375"><figcaption></figcaption></figure>
{% endstep %}

{% step %}

#### Select GPU and name the server

1. Set **Hardware accelerator** to **GPU**
2. Choose a GPU type (for example **T4**, if available)
3. Give the server a name (anything you like)

<figure><img src="https://3215535692-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FxhOjnexMCB3dmuQFQ2Zq%2Fuploads%2FLuK8oheihdCNn9MAZSlk%2Fcolab_img_5.png?alt=media&#x26;token=e5ed396f-afe2-4c09-b1b8-120c7301793c" alt="" width="375"><figcaption></figcaption></figure>

{% hint style="info" %}
Note: GPU availability depends on your Colab plan and current capacity. If you don’t see GPU options, see troubleshooting below.
{% endhint %}
{% endstep %}

{% step %}

#### Pick the Python kernel

Once connected to the Colab server, select the **Python** kernel that appears for that runtime (usually a Python 3 kernel).

<figure><img src="https://3215535692-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FxhOjnexMCB3dmuQFQ2Zq%2Fuploads%2FBexUiXJ7LRFBXs95tX2S%2Fcolab_img_6.png?alt=media&#x26;token=391e343d-d2fa-4a77-8e3f-ef244adfcb4c" alt="" width="375"><figcaption></figcaption></figure>
{% endstep %}

{% step %}

#### Run the notebook

* Click **Run All** in the notebook toolbar (or run cells top-to-bottom)
* Watch the setup cells install dependencies and then start the Unsloth workflow
* You can view our dedicated [fine-tuning](https://unsloth.ai/docs/get-started/fine-tuning-llms-guide) or [reinforcement learning](https://unsloth.ai/docs/get-started/reinforcement-learning-rl-guide) guides for more info on exactly how to get started with Unsloth.
  {% endstep %}
  {% endstepper %}

### Video Tutorial

{% embed url="<https://youtu.be/5NNggTrEWNQ>" %}

### Troubleshooting

#### After Colab server disconnects, the notebook won’t run on a new server

**What’s happening:**\
If the notebook stays open while the Colab server disconnects, VS Code can get stuck in a bad kernel/runtime state after reconnecting. Related [GitHub issue](https://github.com/googlecolab/colab-vscode/issues/200).

**Fix:** Close the notebook tab completely and open the notebook again.

#### You can’t select a GPU (only CPU shows up)

Possible causes and fixes:

* **Colab free tier capacity:** GPUs may be temporarily unavailable → try again later.
* **Not actually connected to a Colab runtime:** re-check **Select Kernel → Colab** and ensure a Colab server is active.
* **Account/region restrictions or limits reached:** you may need to wait or use a different Google account / plan.

#### Everything worked, but packages are “gone” after reconnecting

Colab runtimes are **ephemeral**. When a server restarts, you usually need to re-run the setup/install cells (often the first few cells in the notebook).