equationl
commented
2 months ago Using another model and try again, the recognition effect is more affected by the model.
Closed cagnulein closed 1 month ago
equationl
commented
2 months ago Using another model and try again, the recognition effect is more affected by the model.
cagnulein
commented
2 months ago hi @equationl thanks for the reply. could you please be more specific?
equationl
commented
1 month ago Here is my recognition result:
0.
0:22
30:00
0
2
kilometers
elapsed
pace/km
cals (est)
vertgain（m)
LAP TIME
12
0:22
16
LAP1
3
%
14
COMPLETE
10
8
12
10
。
9
AU
400 meters
tracklength
2
0.0
2.0
8
I
Workout
R
incline
FIT>
speedboth incline 0.0 and speed 2.0 recognized.
May you could try change model at OCR Model List .
The above results are what I got using the Chinese Model
cagnulein
commented
1 month ago thanks @equationl did you use the demo app committed in the repository? or any other version? for my last test i cloned exactly your repo. could you please tell me exactly the models that you used linking them here? Thanks!
equationl
commented
1 month ago Yes, I using this repository's fastdeploydemo module.
The model I used has been uploaded as an attachment
cagnulein
commented
1 month ago Awesome! thanks for your support!
cagnulein
commented
1 month ago sorry @equationl it's my first time with paddle4ocr.
In the demo that I was using there are these 3 files
instead in your zip file I can see these 6 files
how can I replace them? should I run some tools to generate the 3 files?
equationl
commented
1 month ago It seems that you are using paddle-lite and I am using fastDeploy. The model I provide is also a fastDeploy model rather than a paddle-lite model.
Please try running fastdeploydemo from the repository instead of app
cagnulein
commented
1 month ago yes you were right, now it's working! so which are the main differences between the projects?
equationl
commented
1 month ago The two projects are different deployment methods of Paddle.
You can check at FastDeploy and Paddle-Lite
cagnulein
commented
1 month ago Perfect, got it now! Thanks for all the support!
Hi! Thanks for your hard work! Can you please test this screenshot? I can't get all details with the OCR engine with your test app. Am i missing something?
For example the current incline and speed text are both missing
this is the log from the app
let me know if you have any thoughts! thanks!