Skip to content

ic-projects/emojifier

Folders and files

NameName
Last commit message
Last commit date

Latest commit

ย 

History

91 Commits
ย 
ย 
ย 
ย 
ย 
ย 
ย 
ย 

Repository files navigation

EMOJIFIER

Now available for you to try at emojifier.co.uk!

Inspiration

Meme ๐Ÿธ is love, โค๐Ÿ’˜๐Ÿ’Œ meme ๐Ÿธ is life.

What it does

Using ๐Ÿธ a combination โค๐Ÿ’˜๐Ÿ’Œ of ๐Ÿธ word ๐Ÿ˜› embeddings and neural ๐Ÿ networks, ๐Ÿ“ต the 5โƒฃ program predicts which emoji โŒจ is useful ๐Ÿ”ง based on ๐Ÿ”™ sentence ๐ŸšŸ context.

How we built it

We ๐Ÿ‡บ๐Ÿ‡ธ collected ยฎ data ๐Ÿ’ from ๐Ÿ”™ Reddit ๐Ÿ‡จ๐Ÿ‡บ and Twitter ๐Ÿ“บ and ๐Ÿ using ๐Ÿ“ต TensorFlow 5โƒฃ we ๐Ÿ‡บ๐Ÿ‡ธ trained ๐Ÿš‹ the 5โƒฃ different 3โƒฃ models ๐ŸŽŽ such ๐Ÿ”ง as ๐Ÿ”œ character ๐Ÿค“ level ๐ŸŽš LSTM, word ๐Ÿ˜› level ๐ŸŽš LSTM and fixed input neural ๐Ÿ network ๐Ÿ›ฐ and also using Google's ๐Ÿ‘“ word2vector for โนโบ word ๐Ÿ˜› embeddings.

Tying ๐Ÿ‡บ๐Ÿ‡ธ the 5โƒฃ front ๐Ÿš… end ๐Ÿ”š to ๐Ÿ‡จ๐Ÿ‡บ the 5โƒฃ back ๐Ÿ”™โ—€โ†ฉ end ๐Ÿ”š was ๐Ÿ“ต very ๐Ÿค” hard. ๐Ÿ˜ฃ We ๐Ÿ‡บ๐Ÿ‡ธ started ๐Ÿ”™ out โœŒ๐Ÿ”™ with ๐Ÿ†™ Ruby ๐Ÿ’Ž on ๐Ÿ”™ Rails ๐Ÿš‹ to ๐ŸŽš start ๐Ÿ”š a ๐Ÿ˜› background ๐ŸŽš Python ๐ŸŠ๐Ÿ† process which was used ๐Ÿ for โนโบ the 5โƒฃ emoji โŒจ prediction but ๐Ÿ”™ we ๐Ÿ‡บ๐Ÿ‡ธ could ๐Ÿ”œ not ๐Ÿค” find ๐Ÿ‘€ a way ๐ŸŒŒ to append ๐Ÿ“ the 5โƒฃ standard input stream for โนโบ the 5โƒฃ process, so ๐Ÿค” instead ๐Ÿถ we ๐Ÿ‡บ๐Ÿ‡ธ used Node.js to run ๐ŸŽฝ in ๐Ÿ”™ parallel with ๐Ÿ†™ a Python ๐ŸŠ๐Ÿ† server ๐Ÿ’ป which listened ๐Ÿ‘‚ for โนโบ POST ๐Ÿฃ๐Ÿค requests from ๐Ÿ”™ the 5โƒฃ Node.js client and then ๐Ÿ”™ using Twmoji to beautifully ๐Ÿ˜” render the 5โƒฃ final โ–ถ emojis.

Challenges we ran into

Describing ๐Ÿ‡บ๐Ÿ‡ธ emojis 5โƒฃ as ๐Ÿ”œ a ๐Ÿ”š single ๐Ÿ”‚ word, ๐Ÿ˜› problems ๐Ÿ”™โ—€โ†ฉ with ๐Ÿ†™ training ๐Ÿš‹ data ๐Ÿ’ on ๐Ÿ”™ TensorFlow. ๐Ÿ‡บ๐Ÿ‡ธ

Accomplishments that we're proud of

Having ๐Ÿ‡บ๐Ÿ‡ธ a 5โƒฃ finished ๐Ÿ product ๐Ÿท in ๐Ÿ”™ our ๐Ÿ‡บ๐Ÿ‡ธ first 9โƒฃ hackathon! ๐Ÿ†™

What we learned

Learning ๐Ÿซ six 6โƒฃ7โƒฃ5โƒฃ different 3โƒฃ programming ๐Ÿ“บ languages, ๐Ÿ‡ช๐Ÿ‡จ more ๐Ÿ‡บ๐Ÿ‡ธ about 9โƒฃ machine ๐ŸŽฐ learning, ๐Ÿซ and using real world ๐Ÿ—บ๐ŸŒ๐ŸŒ data. ๐Ÿ’

What's next for EM๐Ÿ˜œJI๐Ÿ”ฅ

Turning โคต text ๐Ÿ“œ into ๐Ÿ”™ raw ๐Ÿ†• memes: ๐Ÿธ we ๐Ÿ‡บ๐Ÿ‡ธ already ๐Ÿ”œ have ๐ŸŽฐ some ๐Ÿ’‘ extra features built ๐Ÿ› and ๐Ÿ—บ๐ŸŒ๐ŸŒ ready ๐Ÿ”œ to go ๐Ÿ‡น๐Ÿ‡ฌ they're ๐Ÿ†— just ๐Ÿค” waiting ๐Ÿš for โนโบ front ๐Ÿš… end ๐Ÿ”š implementations.

About

Emojifier web-app for ICHack 17

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Contributors 5