Closed kanakmi closed 3 years ago
@kanakmi i believe you missed out the model itself. All i see is the readme file and the jupyter notebook.
@akshitadixit The trained model is around 300 mb in size which exceeds the maximum single file size that GitHub can support. The Jupyter Notebook contains all the steps required to train the model so users would have to download the dataset and train the model on their local system.
@Isha307 @Shady2320 drop your suggestions please.
@Isha307 @Shady2320 drop your suggestions please.
We are creating the website where user can detect different diseases, So how can we ask user to do all these steps and create your own model and then check yourself.
@Isha307 In that case I can upload the trained model on the Drive and upload the link in the Readme file. Along with that, I can create a simple python script that would accept the image path as an argument and load the model from the disk, make the predictions and return the predictions in a human-readable format. Can this work?
@Isha307 In that case I can upload the trained model on the Drive and upload the link in the Readme file. Along with that, I can create a simple python script that would accept the image path as an argument and load the model from the disk, make the predictions and return the predictions in a human-readable format. Can this work?
Sounds good to me.
@Isha307 @akshitadixit I have made the required changes, please review them.
It's been quite some time since the last message, I am a bit anxious. Is everything OK with the PR or do I have to make some more changes?
Hey @kanakmi I thought @Isha307 would be taking this forward so I laid back on this. Don't worry, she might be occupied, I will review. Meanwhile do let me know if utilize.py works just for the local web-app right?
@akshitadixit It can run on the server as well. All it requires is the trained model (that has to be downloaded from the drive) and the required libraries mentioned in the readme file. Currently, we are passing the image from the local path. On the cloud, we can pass either pass the path or URL of the image (depends on the way you allow the user to upload the image) but there would be a small change in the usage of OpenCV in the latter, the rest would remain the same.
All it requires is the trained model (that has to be downloaded from the drive) and the required libraries mentioned in the readme file.
Sure thing. I'll clone your fork to try this out and will be back on this.
Sure thing. I'll clone your fork to try this out and will be back on this. Did it work alright? Do ping me if there is some error or maybe if you need some test images.
This issue has been automatically marked as stale because it has not had recent activity. It will be closed if no further activity occurs.
This is regarding the issue I opened several days back - https://github.com/IIITKalyaniFOSC/MediCare-Prime/issues/33 I have added the promised model with 95% accuracy. All the extended details are visible in the Readme file.