How to Install Specific Versions of Packages in Google Colab (Step-by-Step Guide)

Sharing is caring!

Introduction

Google Colab comes with many preinstalled libraries, but they’re not always the version you need. This often causes compatibility issues, unexpected errors, or broken code. Knowing how to install specific versions of packages in Colab is essential for reproducibility, debugging, and working with older or project-specific dependencies.

In this guide, you’ll learn all the correct and safe ways to install exact package versions in Google Colab, with beginner-friendly steps, real examples, common mistakes, and best practices used by professionals.

Why You Need Specific Package Versions in Colab

Installing a specific version helps you:

• Avoid breaking changes in newer releases
• Match project or tutorial requirements
• Reproduce experiments reliably
• Fix dependency conflicts
• Run legacy code successfully

Colab environments update frequently, so version control matters more than you think.

How Package Installation Works in Google Colab

Google Colab uses:

• Python
• pip as the default package manager
• A temporary runtime (resets after disconnect)

✅ Installed packages last only for the current session.

How to Install a Specific Version of a Package in Colab (Basic Method)

Using pip (Most Common Way)

Use this syntax:

!pip install package_name==version_number

Example: Install a Specific NumPy Version

!pip install numpy==1.23.5

✅ This replaces the existing version for the current session.

How to Check the Installed Package Version

After installation, always verify:

import numpy
numpy.__version__

Or using pip:

!pip show numpy

✅ This avoids silent version mismatch issues.

Installing Multiple Packages With Specific Versions

You can install multiple versions at once:

!pip install numpy==1.23.5 pandas==1.5.3 matplotlib==3.7.1

✅ Useful for project setup.

Installing Specific Versions From a Requirements File

If you have a requirements.txt file:

Example requirements.txt

numpy==1.23.5
pandas==1.5.3
scikit-learn==1.2.2

Install in Colab:

!pip install -r requirements.txt

✅ Best practice for reproducibility.

How to Downgrade a Package in Google Colab

Sometimes Colab comes with a newer version than you need.

Example: Downgrade TensorFlow

!pip install tensorflow==2.11.0

⚠️ Restart runtime after major library changes:

Runtime → Restart runtime

Installing a Specific Version of TensorFlow in Colab

TensorFlow is a common pain point.

Example:

!pip install tensorflow==2.12.0

After installation:

import tensorflow as tf
tf.__version__

✅ Always restart runtime for TensorFlow changes.

Installing Specific Versions of Packages With pip and –no-deps

To avoid dependency conflicts:

!pip install numpy==1.23.5 --no-deps

⚠️ Use this only if you know dependencies are already satisfied.

How to Install a Package Version From GitHub

Useful when a specific version isn’t on PyPI.

!pip install git+https://github.com/user/repo.git@tag_or_commit

Example:

!pip install git+https://github.com/psf/requests.git@v2.28.1

Common Errors and How to Fix Them

Error 1: “Could not find a version that satisfies the requirement”

Causes:

• Version doesn’t exist
• Python version incompatibility

Fix:

• Check available versions:

!pip index versions package_name

Error 2: Package Installs but Code Still Breaks

Fix:

• Restart runtime
• Re-run all cells
• Verify version after restart

Error 3: Dependency Conflicts

Fix:

• Install packages in the correct order
• Use a requirements file
• Avoid mixing pip and conda

Best Practices for Installing Specific Versions in Colab

✅ Install packages at the top of the notebook
✅ Pin versions explicitly
✅ Restart runtime after major installs
✅ Use requirements.txt
✅ Avoid unnecessary upgrades
✅ Document versions in markdown

Alternatives to pip in Google Colab

✅ pip is the safest choice.

Example: Full Setup Cell for a Colab Notebook

!pip install numpy==1.23.5 pandas==1.5.3 scikit-learn==1.2.2

Then:

import numpy, pandas, sklearn
print(numpy.__version__, pandas.__version__, sklearn.__version__)

Conclusion

Knowing how to install specific versions of packages in Colab is essential for stable, reproducible, and professional workflows. Whether you’re following a tutorial, debugging errors, or deploying research code, version control prevents headaches.

Always pin versions, verify installs, and restart the runtime when needed.

👉 Start using version-specific installs today to avoid “it works on my machine” problems.

Frequently Asked Questions (FAQ)

1. How do I install a specific package version in Colab?

Use:

!pip install package==version

2. Do installed packages persist in Google Colab?

No. Packages reset when the runtime disconnects.

3. Why do I need to restart runtime after installing packages?

Some libraries load at startup and won’t update until restart.

4. Can I downgrade a package in Colab?

Yes, by installing an older version with pip.

5. How do I check package versions in Colab?

Use:

package.__version__

6. Can I use conda in Google Colab?

Not officially. pip is recommended.

7. What Python version does Colab use?

It depends on the runtime, usually Python 3.x (check with python --version).

8. How do I avoid dependency conflicts?

Pin versions and use a requirements.txt file.

9. Can I install packages from GitHub?

Yes, using pip with a GitHub URL.

10. What happens if a version is incompatible with Colab?

pip will fail or the package will not work correctly.