genp / cocottributes Goto Github PK

Hi there,

Would you perhaps make the python3 joblib Coco attributes file available, please? The file found here seems to be the python2 joblib file.

Regards
Shane

Hi, In the ReadMe for the repo, it says you have improved models based on PyTorch: "New pytorch models! Significant improvement over 2016 baseline network. See updates in pytorch dir."

I have been looking into the PyTorch dir, but could not find the directions to get PyTorch models or a checkpoint from where to start. Do you mean, one can start from an ImageNet trained model with coco-attributes dataset and that you are essentially not providing pre-trained PyTorch models but instead code to train PyTorch models?

Looking forward to your answer.
Thanks,
Touqeer

Hi, I was wondering if you had any documentation on how to run the classifier provided here: http://cs.brown.edu/~gmpatter/cocottributes/coco_attributes_2016.tar.gz. This link can be found on this page: http://cs.brown.edu/~gmpatter/cocottributes.html. Thank you!

Hey!

I'm having some issues installing pip install -r requirements.txt. After installing some of them i get this error:
/usr/local/opt/openssl/include/openssl/x509.h:751:15: note: previous declaration is here X509_REVOKED *X509_REVOKED_dup(X509_REVOKED *rev); ^ 1 error generated. error: command 'clang' failed with exit status 1

Any idea on how to solve this?
Running on macos sierra
Thanks!

Looking through the annotated samples, I noticed that for the gender and age attributes of the person class, the label noise is extremely high. In the range 20 to 30 percent label errors, see photos attached. The largest sources of errors are:

• There's a lot of photos of Person class where the gender is simply not assigned even when it's clearly visible (~30%).
• There's a lot of photos of females that are wrongly labeled as males (~20%).
• For the age categories (Young, Adult, Old), almost all photos are simply labeled Adult, even when the object in question is clearly either Young or Old. There's substantially more old people without the Old label than old people with the Old label...

I haven't checked the other object categories but i suspect the label noise to be of similar magnitude. It would be nice if it was declared in the Repo that the labels in the data-set suffer from high error rate.

Hi @genp I am wondering what license is COCO Attributes labels under. How about commercial use? Thanks!

I tried to load it using pickle in Python 3.7, an error is raised when loading it using pickle.load:

Traceback (most recent call last):
  File "<input>", line 1, in <module>
UnicodeDecodeError: 'ascii' codec can't decode byte 0xf0 in position 726: ordinal not in range(128)

Then I tried with parameter encoding='iso-8859-1' and then another error is raised:

Traceback (most recent call last):
  File "<input>", line 1, in <module>
_pickle.UnpicklingError: unpickling stack underflow