diff --git a/notebooks/training_models.ipynb b/notebooks/training_models.ipynb
index 43fed78..da60a06 100644
--- a/notebooks/training_models.ipynb
+++ b/notebooks/training_models.ipynb
@@ -95,7 +95,7 @@
     "\n",
     "For negative data, we'll use small, already prepared samples of the [fma-large dataset](https://github.com/mdeff/fma) for music, the [FSD50k dataset](https://zenodo.org/record/4060432#.Y-hA2BzMJhE) for noise, and the [Common Voice 11](https://huggingface.co/datasets/mozilla-foundation/common_voice_11_0) dataset for speech.\n",
     "\n",
-    "The fma-large sample can be downloaded [here](https://f002.backblazeb2.com/file/openwakeword-resources/data/fma_sample.zip), and then extracted into the working director.\n",
+    "The fma-large sample can be downloaded [here](https://f002.backblazeb2.com/file/openwakeword-resources/data/fma_sample.zip), and then extracted into the working directory.\n",
     "\n",
     "The FSD50k sample can be downloaded [here](https://f002.backblazeb2.com/file/openwakeword-resources/data/fsd50k_sample.zip), and then extracted into the working directory.\n",
     "\n",