Update README.md

README.md

@@ -1,9 +1,18 @@
 ---
 license: mit
+language:
+- ru
+- en
+pipeline_tag: audio-to-audio
 ---
+# Summary
+Pretrained RingFormer from [codename's RVC fork](https://github.com/codename0og/codename-rvc-fork-4)  
 
-Early pretraining checkpoint for RingFormer in [codename's RVC fork](https://github.com/codename0og/codename-rvc-fork-4)  
-40khz, spin_v2 embedder, trained with multiscale mel loss  
-training data is 10-15% Russian, the rest is English  
+**Sample rate**: 40000  
+**Embedder**: Spin v2  
+**Mel Loss**: Multi-scale  
+**Speakers**: 225  
+**Dataset length**: 235:31:50, of which 37:26:53 is Russian, rest is English.  
 
-*TODO: better description once fully trained*
+# Graphs
+![graphs](graphs.png)