Support for image generation <start_of_image>
Open
Nobody has claimed this yet.
- Dominant language
- Python
- Stars
- 12.5k
- Forks
- 998
- Avg merge
- 3d 13h
- Merged PRs (30d)
- 10
Description
The new Gemma3 models can receive a photo via the -a syntax, and when prompted to modify an existing photo, it looks like it tries but fails. Obviously showing images on the CLI is out of scope, but it'd be wonderful if they can be saved in the current directory.
> llm -m gemma3:27b "can you add a mohawk to the gnomes in this photo?" -a gnome\ couple.png
Okay, here's the image with mohawks added to the gnomes! I've tried to make them look somewhat natural within the game's style.
<start_of_image> # <---- Output ends here
Contributor guide
First steps
- Read the whole issue, then the project's contributing guide.
- Comment on the issue to say you are picking it up — it saves two people doing the same work.
- Fork the repository and make your change on a branch.
- Open a pull request that references the issue number.
Research direction
Start at the CLI path that handles the -a argument and inspect how the <start_of_image> response is currently processed. Done means a generated image from a Gemma3 image-editing request is saved in the current directory, with behavior verified using the example command.
Written by the indexing model from the issue text.
Assessment
- Domain
- ai, cli
- Issue type
- Feature
- Difficulty
- 4/5
- Estimated time
- 3-5 days
- Activity status
- Stale
- Clarity
- Mostly clear
- Newbie friendliness
- 35/100