diff --git a/README.md b/README.md
index c5f12a5..ce0fa4f 100644
--- a/README.md
+++ b/README.md
@@ -49,7 +49,7 @@ display(image)
 ```
 <img src="https://github.com/kuprel/min-dalle/raw/main/examples/dali_walle.jpg" alt="min-dalle" width="300"/>
 
-If the model is being used interactively (e.g. in a notebook) `generate_image_stream` can be used to generate a stream of images as it the model is decoding.  The detokenizer adds a slight delay for each intermediate image.  Setting `log2_mid_count` to 3 results in a total of `2 ** 3 = 8` generated images.  The only valid values for `log2_mid_count` are 0, 1, 2, 3, 4.
+If the model is being used interactively (e.g. in a notebook) `generate_image_stream` can be used to generate a stream of images as the model is decoding.  The detokenizer adds a slight delay for each image.  Setting `log2_mid_count` to 3 results in a total of `2 ** 3 = 8` generated images.  The only valid values for `log2_mid_count` are 0, 1, 2, 3, 4.
 
 ```python
 image_stream = model.generate_image_stream(
diff --git a/min_dalle/min_dalle.py b/min_dalle/min_dalle.py
index 76b2bd1..3f23878 100644
--- a/min_dalle/min_dalle.py
+++ b/min_dalle/min_dalle.py
@@ -168,6 +168,7 @@ class MinDalle:
         log2_supercondition_factor: int = 3,
         is_verbose: bool = False
     ) -> Iterator[Image.Image]:
+        assert(log2_mid_count in range(5))
         if is_verbose: print("tokenizing text")
         tokens = self.tokenizer.tokenize(text, is_verbose=is_verbose)
         if is_verbose: print("text tokens", tokens)