feat(SentimentAnalyzer): Improve performance #3
Conversation
|
Thanks for a simple library @arafattehsin! I was wondering what data model this is trained on? There is a list here of ones that could be used for other options: |
There was a problem hiding this comment.
Thanks @kamranayub for such a great improvement. I definitely missed this one. I think I used Yelp Dataset to train this along with some other (which I can't recall) but it was picked up from Kaggle.
From your pointer, I just thought that we could also bring all of the reviews in one library but from different models. I have a strategy for that in my mind which I will shortly write in an issue. Please feel free to chip in your thoughts.
| { | ||
| public static SentimentPrediction Predict(string text) | ||
| private static readonly Sentiments instance = new Sentiments(); |
There was a problem hiding this comment.
This needs to be [ThreadStatic] and initialized once per thread
There was a problem hiding this comment.
Thanks, addressed in #5
Changes
Results
Specs: core i7-3770k @ 3.5ghz, 32gb ram