update README

README.md

@@ -37,7 +37,7 @@ license: cc-by-4.0
 [Cartoon Set](https://google.github.io/cartoonset/) is a collection of random, 2D cartoon avatar images. The cartoons vary in 10 artwork categories, 4 color categories, and 4 proportion categories, with a total of ~1013 possible combinations. We provide sets of 10k and 100k randomly chosen cartoons and labeled attributes.
 
 #### Usage
-`cartoonset` provides the images as byte strings, this gives you a bit more flexibility into how to load the data. Here we show 2 ways:
+`cartoonset` provides the images as PNG byte strings, this gives you a bit more flexibility into how to load the data. Here we show 2 ways:
 
 **Using PIL:**
 ```python
@@ -70,7 +70,7 @@ ds = tf.data.Dataset.from_generator(
 )
 
 def process_fn(sample):
-    img = tf.image.decode_png(sample["img_bytes"], channels=channels)
+    img = tf.image.decode_png(sample["img_bytes"], channels=3)
     ...
     return {"img": img}