johnyburd / 3dfeeling Goto Github PK

Object generation uses emotion function to sample a topographic top layer.
Uses a fixed dimension for area of rectangular base.

I know what I have to do

Continuous function to sample height,y , from based on valence, arousal and dominance.
height = f(v,a,d,x)

Have object generation output something sane when only a single sentence is inputted.

• Frontend sends text to backend
• Backend responds with an image (stl)
• Frontend displays image

Fix bug where too many sentences results in too many points and breaks openpyscad viewer.

transform words to feature vector for classification

Change filename scheme to be the 10 least significant digits of systime as a better naming scheme for generated files.

tokens that occur ~1 times are not informative for classification

"not good" -> not-good

Make a method to evaluate the average loss of the bayes classifier

Change STL file generation so that it displays properly.

The current one is bad. We can do better

add support for adding ngram tokens to classifiers.
Have a function to create n length ngrams out of text.
Add ngrams into classifier.

API should return some fail case if the file contents are unable to be processed

Backend implementation using pyscad

Frontend should check the .{extension} of the file and warn the user if it is not recognized. This warning should arise on the input page

When complete, the backend will send graphs to the frontend so that the user will be able to see visuals of how the data changes over time.

Implement async queue to object generation to receive new requests for text analysis

The user will be able to click a button and then select a file that meets the criteria for our text parsing in the backend

As a user, I want my histogram generator to not randomly break during demos

brief description of methods for extracting values from text.

Have object generation listen on a unix socket for new requests

Put some meat on those bones

The backend needs to be able to accept certain types of files and extract text from them.

File types: .txt, .docx?, .pdf (stretch)

period of 2pi

Introduction to project. Maybe this should be on home page? or something similar?

GPT2 for vad classification

explanation on what VAD represents