-
Notifications
You must be signed in to change notification settings - Fork 0
[2022] Prevođenje znakovnog jezika #11
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
base: master
Are you sure you want to change the base?
Conversation
…m/pfe-rs/izvestaji into 2022-prevodjenje-znakovnog-jezika
Deploying with
|
Latest commit: |
4d52668
|
Status: | ✅ Deploy successful! |
Preview URL: | https://09aab703.pfe-izvestaji.pages.dev |
Branch Preview URL: | https://2022-prevodjenje-znakovnog-j.pfe-izvestaji.pages.dev |
…m/pfe-rs/izvestaji into 2022-prevodjenje-znakovnog-jezika
|
||
Korišćenjem EfficientNetB3 dobijen je loss od 0.44 i preciznost od 99%. | ||
|
||
_grafici_ |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Dodajte grafike
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Teskt je manje vise ok. Dodajte samo opise ispod svake slike
|
||
##### Predprocesing | ||
|
||
Svaka slika je pre klasifikacije izmenjena na nekoliko načina. Na svakoj slici su primenjene sledeće transformacije: |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Dopisite zast menjate slike
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Potrebne neke ispravke u formatiranju pre objavljivanja na sajt.
Takođe, nisam ostavio komentar za to, ali odeljak sa rezultatima treba da bude odvojen odeljak iznad zaključka, a ne ovako rezultati ispod svakog metoda kao sad.
Za poboljšanje konvolucionih neuronskih mreža česta je praksa skaliranje po jednoj od sledeće tri dimenzije - dubina, širina i veličina slike. U [[6]](https://arxiv.org/pdf/1905.11946.pdf) predložen je state-of-the-art način skaliranja mreže, u kome se dobija balans ove tri dimenzije skalirajući svaku od njih istom konstantom. Ovako je razvijeno osam mreža EfficientNetB0 do EfficientNetB7. | ||
Radi poklapanja rezolucije slika iz baze i ulaznih podataka mreže implementirana je EfficientNetB3 mreža. | ||
|
||
_arhitektura_ |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Razrešite komentar kad dodate.
Ovde ću da vam pišem komentare