This is a submission for the Hacktoberfest Open-Source AI Challenge Week 1: Touch Grass
What if the best AI experience was one that made you close...
For further actions, you may consider blocking this person and/or reporting abuse
"Five minutes with the app, the next fifty-five with the real world" is the cleanest product sentence I've read this week, and "my friend won, I still counted it as a win" is the only user test in this challenge I fully believe. I turned the post into an 80-second hand-drawn cartoon: the scrolling trap, the quest card, "your phone is no longer needed", the park, and the friend who beat you. No strings, yours to use or ignore, on a plain page with no sign-up:
ssap-pa.github.io/scripttoon-samples/
If any detail is off, say which shot and I'll re-render it. One question: when Gemma scores a photo, does it get the original objective text alongside the image, or only the image?
Ideally gemma should get the image of the user and description. And based on the given quest + image, ai should produce score and confidence in output and accordingly the user should be scored.
That's the right shape: quest text plus the photo in one call, and a score with a confidence instead of a yes/no, so a borderline "three leaf shapes" can be a partial rather than a fail. Thanks for answering, and for the project.
Im happy that you liked the project
As you are using ollama to interact with gemma, it is not available offline... right ?
By offline means, you need to clone the repo and setup ollama + any ai model. I used gemma in my case
Actually, I am asking- can we use this application by turning off the internet ?
Yes, if you setup locally. No, if your are accessing from internet as it would required internet to download the page and logic
Is there a proper guide to follow for gemma model setup ?
You can google it, read readme.md or ask ai to set it up for you. Anything worksπ
This concept is so refreshing. Most AI tools try to hook users for longer screen time, but building an app whose core goal is to get people outside is brilliant. The pluggable AI provider pattern is really clean too, love how you separated the UI logic from the model implementation.
I'm glad that you liked it. You can also check my other articles where I discuss about mental health, tech and much more :)
Γ tΖ°α»ng tαΊ‘o ra mα»t con AI khuyαΊΏn khΓch ngΖ°α»i dΓΉng "rα»i xa mΓ n hΓ¬nh" thα»±c sα»± rαΊ₯t thΓΊ vα» vΓ cΓ³ chΓΊt chΓ’m biαΊΏm sΓ’u sαΊ―c. Trong quΓ‘ trΓ¬nh lΓ m viα»c vα»i cΓ‘c mΓ΄ hΓ¬nh ngΓ΄n ngα»― lα»n, mΓ¬nh thΖ°α»ng xuyΓͺn rΖ‘i vΓ o cΓ‘i bαΊ«y lαΊ‘m dα»₯ng chΓΊng cho mα»i tΓ‘c vα»₯ nhα» nhαΊ₯t, dαΊ«n ΔαΊΏn viα»c mαΊ₯t Δi khαΊ£ nΔng tΖ° duy Δα»c lαΊp. Viα»c thiαΊΏt kαΊΏ mα»t agent cΓ³ "Γ½ thα»©c" vα» sα»± phα»₯ thuα»c của con ngΖ°α»i lΓ mα»t hΖ°α»ng Δi rαΊ₯t nhΓ’n vΔn cho cΓ‘c dα»± Γ‘n open-source. MΓ¬nh cΕ©ng Δang thα» nghiα»m cΓ‘c workflow tα»± Δα»ng hΓ³a thΓ΄ng minh hΖ‘n Δα» tiαΊΏt kiα»m thα»i gian thay vΓ¬ chα» ngα»i chat liΓͺn tα»₯c, mΓ¬nh tΓ¬m thαΊ₯y vΓ i tΓ i liα»u hay vα» viα»c quαΊ£n lΓ½ quota mΓ΄ hΓ¬nh qua LabAgent, site: labagent .tech Δα» tα»i Ζ°u hΓ³a quy trΓ¬nh mΓ khΓ΄ng bα» quΓ‘ tαΊ£i.
That's really great. Please share your findings here and let's deep dive.
Great idea, and the GIFs in the article are hilarious too
Thanks, I like the GIFs tooππ
I'm glad that you liked the post. Please share your feedback and let's hear what are you cooking this time!