feat(blogs-corpus): new corpus for word2vec

src/Sibe.hs

@@ -26,6 +26,7 @@ module Sibe
      sigmoid',
      softmax,
      softmax',
+     sampledSoftmax,
      relu,
      relu',
      crossEntropy,
@@ -181,6 +182,12 @@ module Sibe
         where
           sig x = 1 / max (1 + exp (-x)) 1e-10
 
+      -- used for negative sampling
+      sampledSoftmax :: Int -> Vector Double -> Vector Double
+      sampledSoftmax n x = cmap (\a -> exp a / s) x
+        where
+          s = V.sum . exp $ V.take n x
+
       relu :: Vector Double -> Vector Double
       relu = cmap (max 0.1)
 

src/Sibe/Word2Vec.hs

@@ -14,6 +14,7 @@ module Sibe.Word2Vec
     import Data.Default.Class
     import Data.Function (on)
     import Control.Monad
+    import System.Random
 
     data W2VMethod = SkipGram | CBOW
     data Word2Vec = Word2Vec { docs :: [String]
@@ -27,8 +28,9 @@ module Sibe.Word2Vec
                      }
 
     word2vec w2v session = do
+      seed <- newStdGen
       let s = session { training = trainingData
-                      , network = randomNetwork 0 (-1, 1) v [(dimensions w2v, (id, one))] (v, (softmax, one))
+                      , network = randomNetwork 0 (-1, 1) v [(dimensions w2v, (id, one))] (v, (softmax, crossEntropy'))
                       }
 
       when (debug s) $ do