diff --git a/README.md b/README.md
index 25d595a..24e2404 100644
--- a/README.md
+++ b/README.md
@@ -457,7 +457,7 @@ tuple(tuple(a,b), noise)
 
 Where `a` is the input sample, `b` is the label/condition (if any, otherwise fill it with `0`), and `noise` is the latent vector of input.
 
-To train Pix2Pix-like architecture, that have no `noise` as ConvGenerator input, just return the values in thee format `(tuple(a,b), b)` since the condition is the generator output.
+To train Pix2Pix-like architecture, that have no `noise` as ConvGenerator input, just return the values in thee format `(tuple(a,b), b)` since the condition is the generator input.
 
 ## Test
 In order to run the tests (with the doctests), linting and docs generation simply use `tox`.
diff --git a/src/ashpy/models/convolutional/unet.py b/src/ashpy/models/convolutional/unet.py
index 8de13a9..3b464c7 100644
--- a/src/ashpy/models/convolutional/unet.py
+++ b/src/ashpy/models/convolutional/unet.py
@@ -27,7 +27,10 @@ class UNet(Conv2DInterface):
     """
     UNet Architecture.
 
-    Used in Image-to-Image Translation with Conditional Adversarial Nets [1]_.
+    Architecture similar to the one found in "Image-to-Image Translation
+    with Conditional Adversarial Nets" [1]_.
+
+    Originally proposed in "U-Net: Convolutional Networks for Biomedical Image Segmentation" [2]_.
 
     Examples:
         * Direct Usage:
@@ -50,8 +53,10 @@ class UNet(Conv2DInterface):
                 (1, 512, 512, 3)
                 True
 
-    .. [1] Image-to-Image Translation with Conditional Adversarial Nets
-        https://arxiv.org/abs/1611.04076
+    .. [1] Image-to-Image Translation with Conditional Adversarial Nets -
+        https://arxiv.org/abs/1611.07004
+    .. [2] U-Net: Convolutional Networks for Biomedical Image Segmentation -
+        https://arxiv.org/abs/1505.04597
 
     """