Comments (12)
Hey, Sorry for the delayed reply. I have been sick for a few days. Will solve the bugs in the master by the end of the day. Thank You for your patience.
from craft-remade.
from craft-remade.
How about the recall and precision on ICDAR2013 by your implemention?
The weak supervision code seems to be working. I have updated the master branch to incorporate the changes. I will train the model tomorrow on IC13, IC15 and send the link to the pre-trained model, Generated character Bounding Boxes, and F-score.
Thank You for your patience.
from craft-remade.
Hi @mayank-git-hub , thanks for your code. Do you mind telling me that what is the performance of the pre-trained model on ICDAR 13 dataset?
from craft-remade.
Currently the performance is not good. I am able to achieve an approximate F-score of 70. I am tweaking and experimenting to improve the results. Will update the ReadMe file with the F-scores of all the datasets.
from craft-remade.
Thanks for your reply. I am also working on the training code. And also only could get 0.71 H-mean on IC13 dataset. Looking for your results.
from craft-remade.
Great, I am also working on it. Would update the Readme file as soon as I get better results. Thank You for your patience.
from craft-remade.
@mayank-git-hub Quick question, 0.7051 F1 Score on IC13 is with CRAFT or ResNet CRAFT ? Which one is better and by how much ?
Thanks ;)
from craft-remade.
The f score is with craft model with vgg backbone. Haven't tried resnet unet architecture.
from craft-remade.
Thanks @mayank-git-hub. Btw, I saw you included link to Original Model by authors in the README. Can we load that model and fine tune it with this code ?
from craft-remade.
Thanks @mayank-git-hub. Btw, I saw you included link to Original Model by authors in the README. Can we load that model and fine tune it with this code ?
Yep we can ;)
from craft-remade.
Yep
from craft-remade.
Related Issues (20)
- How to load pkl model with torch ? HOT 1
- utils_old.py HOT 4
- RuntimeError: DataLoader worker (pid 4249) is killed by signal: Killed. HOT 1
- License HOT 1
- CUDA memory consumption HOT 1
- Error on loading final_model.pkl HOT 1
- Weak Supervision Training HOT 1
- Model got after training HOT 10
- SynthText dataset HOT 2
- Error: No such command 'train_synth'. HOT 4
- Paths in a train_synth/config.py point to a Google drive directory that is not shared. HOT 2
- This is right? (weight file) HOT 2
- How to reproduce the provided ICDAR2013 model? HOT 1
- Loading cache.pkl while debugging HOT 1
- gt.mat file HOT 1
- No Output After Training on Weak Supervision HOT 2
- strong train path HOT 1
- how can I get gt.mat?
- When I run'python main.py synthesize --model=./model/63000_model.pkl --folder=./input', I got 'Segmentation fault (core dumped)'
- Suggestions needed for custom training
Recommend Projects
-
React
A declarative, efficient, and flexible JavaScript library for building user interfaces.
-
Vue.js
🖖 Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.
-
Typescript
TypeScript is a superset of JavaScript that compiles to clean JavaScript output.
-
TensorFlow
An Open Source Machine Learning Framework for Everyone
-
Django
The Web framework for perfectionists with deadlines.
-
Laravel
A PHP framework for web artisans
-
D3
Bring data to life with SVG, Canvas and HTML. 📊📈🎉
-
Recommend Topics
-
javascript
JavaScript (JS) is a lightweight interpreted programming language with first-class functions.
-
web
Some thing interesting about web. New door for the world.
-
server
A server is a program made to process requests and deliver data to clients.
-
Machine learning
Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.
-
Visualization
Some thing interesting about visualization, use data art
-
Game
Some thing interesting about game, make everyone happy.
Recommend Org
-
Facebook
We are working to build community through open source technology. NB: members must have two-factor auth.
-
Microsoft
Open source projects and samples from Microsoft.
-
Google
Google ❤️ Open Source for everyone.
-
Alibaba
Alibaba Open Source for everyone
-
D3
Data-Driven Documents codes.
-
Tencent
China tencent open source team.
from craft-remade.