Created model card for first successfull training.

Hopefully now that I got it training with the dataset I can try again training the model.

README.md

@@ -0,0 +1,26 @@
+
+---
+language: nah specifically ncj
+datasets:
+- created a new dataset based on https://www.openslr.org/92/
+metrics:
+- wer
+tags:
+- audio
+- automatic-speech-recognition
+- speech
+- xlsr-fine-tuning-week
+license: apache-2.0
+model-index:
+- name: Nahuatl XLSR Wav2Vec 53
+  results:
+  - task: 
+      name: Speech Recognition
+      type: automatic-speech-recognition
+    metrics:
+       - name: Test WER
+         type: wer
+         value: 69.11
+---
+
+First full try on the new dataset. Updates to come.