-
Notifications
You must be signed in to change notification settings - Fork 128
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Add Count tools #56
Add Count tools #56
Conversation
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
looks good, few minor comments
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Forgot, make sure to add the new tools to the README tools section
…zed coordinates, refactoring code, adding llm generate counter tool
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Looks good! left one optional comment. I think it's just failing tests because of linting so can merge once that's fixed.
A dictionary containing the key 'count' and the count as value. E.g. {count: 12} | ||
""" | ||
image_size = get_image_size(image) | ||
bbox = [float(x) for x in prompt.split(",")] |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
This is minor, but I would just require the bounding boxes as input instead of a prompt so it doesn't have to worry about parsing.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Sounds good, i would make it generic so that it can be bbox or prompt. Will include it in the next PR
* Adding counting tools to vision agent * fixed heatmap overlay and addressesessed PR comments * adding the counting tool to take both absolute coordinate and normalized coordinates, refactoring code, adding llm generate counter tool * fix linting
* added custom tools * updated readme * register tool returns tool' * Add a new tool: determine if a bbox is contained within another bbox (#59) * Add a new bounding box contains tool * Fix format * [skip ci] chore(release): vision-agent 0.1.5 * Add Count tools (#56) * Adding counting tools to vision agent * fixed heatmap overlay and addressesessed PR comments * adding the counting tool to take both absolute coordinate and normalized coordinates, refactoring code, adding llm generate counter tool * fix linting * Remove torch and cuda dependencies (#60) Resolve merge conflicts * [skip ci] chore(release): vision-agent 0.2.1 * make it easier to use custom tools * ran isort * fix linting error * added OCR * added example template matching use case * formatting and typing fix * round scores * fix readme typo --------- Co-authored-by: Asia <[email protected]> Co-authored-by: GitHub Actions Bot <[email protected]> Co-authored-by: Shankar <[email protected]>
This PR adds 2 tools to the vision-agent