From a66405bdca348a6e0c480fa998fd64b005f52da0 Mon Sep 17 00:00:00 2001
From: Loreto Parisi <loretoparisi@gmail.com>
Date: Tue, 17 Oct 2017 14:31:44 +0200
Subject: [PATCH 1/3] Updated README with virtualenv install and emoji codes
 mapping

---
 README.md        | 13 +++++++++++++
 requirements.txt |  7 +++++++
 2 files changed, 20 insertions(+)
 create mode 100644 requirements.txt

diff --git a/README.md b/README.md
index 2ca957a..efd42b5 100644
--- a/README.md
+++ b/README.md
@@ -45,6 +45,14 @@ Run the included script, which downloads the pretrained DeepMoji weights (~85MB)
 python scripts/download_weights.py
 ```
 
+## Virtualenv installation with Default backend (Tensorflow)
+To install via `virtualenv` and the defaulk leras backend `tensorflow`
+```
+virtualenv .env
+. .env/bin/activate
+pip install -r requirements.text
+```
+
 ## Testing
 To run the tests, install [nose](http://nose.readthedocs.io/en/latest/). After installing, navigate to the [tests/](tests) directory and run:
 
@@ -52,6 +60,11 @@ To run the tests, install [nose](http://nose.readthedocs.io/en/latest/). After i
 nosetests -v
 ```
 
+## Emoji Codes
+Here the moji codes mapping.
+
+![emoji_overview](https://user-images.githubusercontent.com/163333/31664953-a4ea8836-b347-11e7-82b4-999b41949786.png)
+
 By default, this will also run finetuning tests. These tests train the model for one epoch and then check the resulting accuracy, which may take several minutes to finish. If you'd prefer to exclude those, run the following instead: 
 
 ```bash
diff --git a/requirements.txt b/requirements.txt
new file mode 100644
index 0000000..d0864c0
--- /dev/null
+++ b/requirements.txt
@@ -0,0 +1,7 @@
+tensorflow
+emoji==0.4.5
+h5py==2.7.0
+Keras==2.0.7
+numpy==1.13.1
+scikit-learn==0.19.0
+text-unidecode==1.0

From c053b885ec358abd45ababa46ea1144da745c405 Mon Sep 17 00:00:00 2001
From: Loreto Parisi <loretoparisi@gmail.com>
Date: Tue, 17 Oct 2017 14:33:05 +0200
Subject: [PATCH 2/3] Updated README with virtualenv install and emoji codes
 mapping

---
 README.md | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/README.md b/README.md
index efd42b5..c2ce3eb 100644
--- a/README.md
+++ b/README.md
@@ -45,8 +45,8 @@ Run the included script, which downloads the pretrained DeepMoji weights (~85MB)
 python scripts/download_weights.py
 ```
 
-## Virtualenv installation with Default backend (Tensorflow)
-To install via `virtualenv` and the defaulk leras backend `tensorflow`
+## Virtualenv installation with Default backend
+To install via `virtualenv` and the default Keras backend `tensorflow`
 ```
 virtualenv .env
 . .env/bin/activate

From 766759c48427e3d8b5c6417b1f547d7264936720 Mon Sep 17 00:00:00 2001
From: Loreto Parisi <loretoparisi@gmail.com>
Date: Tue, 17 Oct 2017 14:34:22 +0200
Subject: [PATCH 3/3] Updated README with virtualenv install and emoji codes
 mapping

---
 README.md | 10 +++++-----
 1 file changed, 5 insertions(+), 5 deletions(-)

diff --git a/README.md b/README.md
index c2ce3eb..8807c5b 100644
--- a/README.md
+++ b/README.md
@@ -19,6 +19,11 @@ To start out with, have a look inside the [examples/](examples) directory. See [
 
 Please consider citing our [paper](https://arxiv.org/abs/1708.00524) if you use our model or code (see below for citation).
 
+## Emoji Codes
+Here the emoji codes mapping to model classes.
+
+![emoji_overview](https://user-images.githubusercontent.com/163333/31664953-a4ea8836-b347-11e7-82b4-999b41949786.png)
+
 ## Frameworks
 
 This code is based on Keras, which requires either Theano or Tensorflow as the backend. If you would rather use pyTorch there's an implementation available [here](https://github.com/huggingface/torchMoji), which has kindly been provided by Thomas Wolf.
@@ -60,11 +65,6 @@ To run the tests, install [nose](http://nose.readthedocs.io/en/latest/). After i
 nosetests -v
 ```
 
-## Emoji Codes
-Here the moji codes mapping.
-
-![emoji_overview](https://user-images.githubusercontent.com/163333/31664953-a4ea8836-b347-11e7-82b4-999b41949786.png)
-
 By default, this will also run finetuning tests. These tests train the model for one epoch and then check the resulting accuracy, which may take several minutes to finish. If you'd prefer to exclude those, run the following instead: 
 
 ```bash