feat(todo): re-train model

[QIN2DIM]

In extreme tests, the effect of the following model does not meet the expectations, and we plan to release them in the future.

• [ ] living_room
• [x] bedroom
• [ ] smiling_dog
• [x] conference_room

[M-Zubair10]

Did it updated? How will I know that project is updated? and download new models object.yaml automatically, currently I do manually

[QIN2DIM]

emmm... Let me show you a sketch I made earlier.😂

[QIN2DIM]

For the tags in objects.yaml, Challenger maintains an Assets object that will help you automatically update and replace the model (file of the same name) and automatically download missing models locally as the challenge progresses. You do not need to download all ONNX models from Releases in advance.

[QIN2DIM]

I realized I didn't seem to have finished it.👀👀👀👀👀👀👀👀

[M-Zubair10]

These labels now came rarely, they are updating constantly new prompts, currently bottle broken is hot one Did you solve the fit bounding box around object prompt you said you have solved it gracefully, can you show it

[M-Zubair10]

Can you train object tractor for me? its the only object left for recaptcha solver

[device-id]

Can you train object tractor for me? its the only object left for recaptcha solver

This doesn't have to do with hcaptcha but if you need to train a object detection model on the recaptcha tractors what you can do is just use images of tractors on google and then train it on those. Or if you don't want to do any of that just use a audio solver

[QIN2DIM]

Yes, it's as easy as homework.😂

[M-Zubair10]

I'll definitely try it out soon

On Tue, Sep 20, 2022, 9:11 PM QIN2DIM @.***> wrote:

Yes, it's as easy as homework.😂

— Reply to this email directly, view it on GitHub https://github.com/QIN2DIM/hcaptcha-challenger/issues/157#issuecomment-1252582961, or unsubscribe https://github.com/notifications/unsubscribe-auth/ALL7Y5AXYZHAPQX7U2ZDM6LV7HO2RANCNFSM573UX5IA . You are receiving this because you commented.Message ID: @.***>