Merge branch 'master' of https://github.com/williamleif/histwords

Merging in repo.

README.md

@@ -1,4 +1,4 @@
-#Word Embeddings for Historical Text
+# Word Embeddings for Historical Text
 
 ### Author: William Hamilton (wleif@stanford.edu)
 ### [Project Website](http://nlp.stanford.edu/projects/histwords)

sgns/postprocesssgns.py

@@ -35,7 +35,7 @@ def worker(proc_num, queue, dir, count_dir, min_count):
         write_pickle(iw, dir + str(year) + "-vocab.pkl")
 
 if __name__ == "__main__":
-    parser = ArgumentParser("Constructs training files for GloVe.")
+    parser = ArgumentParser("Post-processes SGNS vectors to easier-to-use format. Removes infrequent words.")
     parser.add_argument("dir")
     parser.add_argument("count_dir", help="Directory with count data.")
     parser.add_argument("--workers", type=int, help="Number of processes to spawn", default=20)

sgns/runword2vec.py

@@ -38,7 +38,7 @@ def train_years(years, in_dir, out_dir, dim, workers, sequential):
                     '-verbose', '2'])
 
 if __name__ == "__main__":
-    parser = ArgumentParser("Runs sequential Glove embeddings for years")
+    parser = ArgumentParser("Runs sequential word2vec embeddings for years")
     parser.add_argument("in_dir", help="Directory with cooccurrence information and vocab.")
     parser.add_argument("out_dir")
     parser.add_argument("--dim", type=int, default=300)