Comments (11)
hi,
the train list is just txt with the image paths.
each image in the should have corresponding txt file, where each line corresponds to word.
line example:
0 0.851475423177 0.247500450534 0.128710019264 0.0338691623536 -0.102781031299 CWRU
cls x y w h alpha txt
cls - class - not used
x - bbox center.x normalized by image width
y - bbox center.y normalized by image height
w - bbox width normalized by image diagonal
h - bbox height normalized by image diagonal
alpha - box angle in radians
from deeptextspotter.
Thanks for your quick response! It's very clearly now.
from deeptextspotter.
Where do we put the file's name?
from deeptextspotter.
Hi we tried train.py but are getting following error:
I0213 04:00:44.575316 17258 net.cpp:242] This network produces output softmax
I0213 04:00:44.575346 17258 net.cpp:255] Network initialization done.
I0213 04:00:44.575435 17258 solver.cpp:56] Solver scaffolding done.
[416, 416]
I0213 04:00:44.581261 17258 solver.cpp:330] Iteration 0, Testing net (#0)
Floating point exception (core dumped)
We are creating a separate .txt file for each .jpeg file. For example, one of the .txt file looks like following:-
0 86.544325 113.70325 61.5000988376 10.8000627304 3.03217953614 calliberation
Please help us resolving the error.
from deeptextspotter.
thanks, I got it, I was not normalizing the coordinates and another mistake, the angle calculation was using the right-handed coordinate system, whereas the image follows the left-handed. That's why my angle was 3 radians.
Also, it's not mentioned here but I am giving image path as the first word in the line of train_list.txt. Hope I am right in that?
from deeptextspotter.
file names are to be put in train_list.txt and valid_list.txt. Each file with image_name.txt in foder "img" wherever your training image folder is present.
from deeptextspotter.
format: cls x y w h alpha txt
Can we give spaces in txt?
from deeptextspotter.
hi,
the train list is just txt with the image paths.
each image in the should have corresponding txt file, where each line corresponds to word.line example:
0 0.851475423177 0.247500450534 0.128710019264 0.0338691623536 -0.102781031299 CWRU
cls x y w h alpha txtcls - class - not used
x - bbox center.x normalized by image width
y - bbox center.y normalized by image height
w - bbox width normalized by image diagonal
h - bbox height normalized by image diagonal
alpha - box angle in radians
Hi, MichalBusta, so we don't use class, just x,y,w,h, alpha and txt?
from deeptextspotter.
hi,
the train list is just txt with the image paths.
each image in the should have corresponding txt file, where each line corresponds to word.
line example:
0 0.851475423177 0.247500450534 0.128710019264 0.0338691623536 -0.102781031299 CWRU
Hi, MichalBusta, so we don't use class, just x,y,w,h, alpha and txt?
Hi, yes, class has been used to predict language of the text (but it did not not work so well, also we think that the language identification step is not needed: https://arxiv.org/abs/1801.09919)
from deeptextspotter.
from deeptextspotter.
t
hi ,I just want ask the single txt file is just the list with x,y,w,h and alpha?I tried that before, but its didn't work . the error is RuntimeError: bad lexical cast:source type value could not be interpreted as target. Could you give me some advise?
from deeptextspotter.
Related Issues (20)
- Training problem HOT 1
- Rotation angle ground truth and image resizing HOT 2
- pre-training ctc_loss
- What do I prepare for re-train ? HOT 1
- CTC model not clearing memory between samples? HOT 2
- I always meet this problem when building Caffe HOT 4
- what is the purpose of buckets? HOT 1
- size mismatch while training HOT 3
- Test result tuning for NVIDIA
- re-train problem
- Train problem HOT 2
- how to pre-train the detection CNN and the recognition CNN
- How long do I have to train? HOT 2
- retrain failed HOT 3
- only RPN are learned
- > hi,
- Extract cordinates and print words as key value pair
- Extract cordinates and print words as key value pair
- Cant Pre-train
- Fine tuning for custom dataset ? HOT 1
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 deeptextspotter.