chen0040 / keras-english-resume-parser-and-analyzer

keras project that parses and analyze english resumes
MIT License
272 stars 144 forks source link

Scroll bar missing in training DL gui #5

Open vasudha-s opened 6 years ago

vasudha-s commented 6 years ago

Unable to scroll down in the GUI that is displayed for the DL based training. Scroll bar is not displayed

aamruta commented 6 years ago

Is this resolved?

ashoknar commented 6 years ago

create_training_data (copy).zip Was stuck with this for a few days. Use this modified file where I've modified it to include a scrollbar.

vasudha-s commented 6 years ago

Hi Ashok,

Thank you! I have a query. I'm caught up with a task for a while now where the training needs to happen automatically. This code helps a lot except except for the UI . Is there a way to train this program without UI or user intervention. I mean, just the feeding of documents should train the parser.

Thanks, Vasudha

On Tue 17 Jul, 2018, 11:22 AM Ashok Narendranath, notifications@github.com wrote:

create_training_data (copy).zip https://github.com/chen0040/keras-english-resume-parser-and-analyzer/files/2200323/create_training_data.copy.zip Was stuck with this for a few days. Use this modified file with a new UX.

— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub https://github.com/chen0040/keras-english-resume-parser-and-analyzer/issues/5#issuecomment-405467488, or mute the thread https://github.com/notifications/unsubscribe-auth/AK6M4Cfi8LhfKbVaeDRjPLz3mTvpWs_Eks5uHXuCgaJpZM4U-QqE .

ashoknar commented 6 years ago

Honestly I had similar thoughts, but when you think about it isn't that what the trained parser does to a document? After it is trained it can accurately classify each line based on the set of line_labels and line_types. If anything but a human labels the data, the model becomes limited to what that entity can identify. Which would make the whole training exercise useless as it would just predict what the entity did in the first place. Wouldn't get any better.

For this to be trained without user intervention I'm guessing some kind of unsupervised clustering model should be implemented. I'm currently not knowledgeable enough on how that would be implemented or what parameters should be adjusted to improve the accuracy of the prediction. I'm also unsure on how a clustering model would be labeled to achieve the required classification.

singhania1408 commented 6 years ago

@ashoknar By using your create_training_data file , it enables scrolling but disables columns of line label and line type options selection. Can you please help out on this?

ashoknar commented 6 years ago

@singhania1408 resize the window or try maximizing it. The columns for line labels and line types are present, don't think my code is responsive to different resolutions. That could be why you can't see it.

jpsilvashy commented 6 years ago

I also have some UI problems, I'd love to help fix this but am unfamiliar with this UI tool, here is a pic of the TKinter UI:

img

Saiteja6034 commented 5 years ago

@ashoknar By using your create_training_data file , it enables scrolling but disables columns of line label and line type options selection. Can you please help out on this?

Try to change parameters in the create_window method..it worked for me

vijender412 commented 5 years ago

@ashoknar By using your create_training_data file , it enables scrolling but disables columns of line label and line type options selection. Can you please help out on this?

Try to change parameters in the create_window method..it worked for me

Hey Saiteja, What exactly are the parameters values. Please share

sayalraza commented 4 years ago

@Saiteja6034 What are the parameters to be changed? Could you share your code?

shubham-mhaske commented 4 years ago

Finally found it guys.

temo.txt

Creative-Priyamvada commented 3 years ago

create_training_data.zip added scrollbar in tkinker gui