Add yelp and yahoo data

OpenVLG · Aug 20, 2022 · 9d06cde · 9d06cde
1 parent f1668b5
commit 9d06cde
Show file tree

Hide file tree

Showing 8 changed files with 260,021 additions and 0 deletions.
diff --git a/data/yahoo/vocab.txt b/data/yahoo/vocab.txt
diff --git a/data/yahoo/yahoo.test.txt b/data/yahoo/yahoo.test.txt
diff --git a/data/yahoo/yahoo.train.txt b/data/yahoo/yahoo.train.txt
diff --git a/data/yahoo/yahoo.valid.txt b/data/yahoo/yahoo.valid.txt
diff --git a/data/yelp/readme.txt b/data/yelp/readme.txt
@@ -0,0 +1,24 @@
+Yelp15 data set used in the following paper:
+
+@article{yang2017improved,
+  title={Improved Variational Autoencoders for Text Modeling using Dilated
+  Convolutions},
+    author={Yang, Zichao and Hu, Zhiting and Salakhutdinov, Ruslan and
+    Berg-Kirkpatrick, Taylor},
+    journal={arXiv preprint arXiv:1702.08139},
+    year={2017}
+    }
+
+DESCRIPTION
+
+This data set is constructed from Yelp15 from:
+Duyu Tang, Bing Qin, Ting Liu.
+Document Modeling with Gated Recurrent Neural Network
+for Sentiment Classification. EMNLP 2017.
+http://ir.hit.edu.cn/~dytang/paper/emnlp2015/emnlp-2015-data.7z
+
+The original dataset contains 1.2 milliion training samples and 156k validation
+and testing samples. We sample 100k as training, 10k as validation and 10k as
+testing from the respective sets.
+
+We use a vocabulary size of 20k and replace out of vocabulary tokens as _UNK.
diff --git a/data/yelp/yelp.test.txt b/data/yelp/yelp.test.txt
diff --git a/data/yelp/yelp.train.txt b/data/yelp/yelp.train.txt
diff --git a/data/yelp/yelp.valid.txt b/data/yelp/yelp.valid.txt