Upload README.md with huggingface_hub

README.md

@@ -1,16 +1,20 @@
 ---
 license: mit
-tags:
-- audio-to-symbolic
-- chart-generation
-- model_hub_mixin
-- music
-- pytorch_model_hub_mixin
-- rhythm-game
-- taiko
+library_name: softchart
+tags: [taiko, beat-tracking, downbeat, music]
 ---
 
-This model has been pushed to the Hub using the [PytorchModelHubMixin](https://huggingface.co/docs/huggingface_hub/package_reference/mixins#huggingface_hub.PyTorchModelHubMixin) integration:
-- Code: https://github.com/JacobLinCool/SoftChart
-- Paper: [More Information Needed]
-- Docs: [More Information Needed]
+# SoftChart Beat
+
+An 8.1M encoder–decoder with a beat/downbeat head, trained on TJA measure grids.
+On held-out songs: **beat F 0.87, downbeat F 0.62** (@70 ms). Used by SoftChart to
+refine BPM and anchor TJA barlines to detected downbeats (handles pickups).
+
+```python
+from softchart.generate import load_hf
+beat_model = load_hf("JacobLinCool/softchart-beat", device="cuda")
+# beat_model.beat(encoder_output) -> (T/4, 2) beat/downbeat logits
+```
+
+Part of [SoftChart](https://github.com/JacobLinCool/SoftChart). MIT-licensed; trained
+on charts derived from a commercial game (research/personal use).