[please] Object detection

Part 1 (2020)
1

hello

is object detection impossible?

I’m trying to make object detection for predicting the new image (not include in train, val, test set)

so, I tried

(1)https://github.com/muellerzr/Practical-Deep-Learning-for-Coders-2.0/blob/master/Computer%20Vision/06_Object_Detection.ipynb

(2)Object detection using fastai v2

(3)fastai v1 (https://github.com/fastai/course-v3/blob/master/nbs/dl2/pascal.ipynb )

but they just predict their validation set … not a unlabeled “new” one.

so i failed above 3 methods…

is there any other method ??
only iceVision possible? i want to make it only by using fastai

sorry for bad English!

thank you!

2

Hey!

If I understand correctly you are looking for the learn.get_preds() or learn.predict() methods.

Once you have your trained model you can then pass it single images or dataloaders of images to it using theses meathods.

Have a look at the Inference functions part of the docs for more info about them

Also you might find Lesson 3 of the course useful - around 18:00 mins Jeremy talks through deploying a model where learn.predict() is used.

Hope that answers things!

2 Likes
3

thank you for reply!!

I tried what you said but it was difficult…(i was able to make classifer but detection was not able to)

and… I tried another thing

so i was able to make predict method but accuracy is very bad.
below one is that one.

i hope my model can predict below labels
two eye_open and one front-face( = 5) in that code picture

could you help me??

i’m doing it with version1 because i can not make predict method on version2