add google colab notebook example

[xprabhudayal]

hey,

this adds a notebook example for running inference on colab. makes it easier to test the model without needing to set up a local env.

tested on t4 gpu (free tier) and works fine with lazy_load enabled.

some output :

1. top of the notebook

---

2. inference

closes #48

[F1shYi]

Dear xprabhudayal, thanks for making this colab tutorial for easier use! It is indeed a convenient way for a quickstart! Would you mind sharing the link to your colab demo? We would link to consider adding the colab link in the README page rather than directly pushing the jupyter notebook file to the main branch.

[F1shYi]

Typo: We would like to consider

[xprabhudayal]

hey, thanks for the fast response.

i totally understand the preference for linking to a hosted colab instead of adding the file directly. however, i think there are a few advantages to having it in the repo:

1. version control: the notebook stays in sync with the codebase. if you update the api or change model paths, the notebook gets updated in the same pr. with an external colab link, it gets outdated pretty fast and becomes a maintenance burden.

2. discoverability: users exploring the examples folder will immediately see the notebook as an option. they don't need to hunt through the readme.

3. reproducibility: the notebook is tied to specific commits. users can checkout an older version of your repo and the notebook will still work with that version's api.

4. contribution friendly: other contributors can easily improve the notebook via pr. with an external colab, only the original creator can update it unless you share edit access.

5. standard practice: most ml repos (huggingface transformers, stable diffusion, flux, etc) ship notebooks in their examples folder for this exact reason.

that said, if you still prefer just a link in the readme, i can host it and share the link. just wanted to make the case for why having it in the repo might be better long term. let me know what you think.