Update README.md

README.md

@@ -1,3 +1,24 @@
----
-license: apache-2.0
----
+---
+license: apache-2.0
+base_model:
+- hexgrad/Kokoro-82M
+pipeline_tag: text-to-speech
+tags:
+- OpenVINO
+- OpenArc
+- CPU
+---
+
+This model was converted using the [kokoro notebook](https://github.com/openvinotoolkit/openvino_notebooks/blob/latest/notebooks/kokoro/kokoro.ipynb) using **weight only compression** to FP16.
+
+Currently Kokoro-82M-FP16-OpenVINO only works on CPU due to an unsupported kernel operation ScatterNDUpdate. Implementing GPU compilation might require writing a custom export openvino_model.xml with the kernel logic to support this. 
+
+After hacking on this unsuccessfully for a few hours, I decided that I should not worry about these details, and use GPU for other tasks. CPU performance is very fast.
+
+
+Required dependencies:
+
+```
+pip install kokoro misaki[en] openvino transformers soundfile 
+```
+