Update README.md

README.md

@@ -1,4 +1,4 @@
-First fine-tuning try of `wav2vec2-large-lv60`. Results are a bit worse than those reported in the Appendix in Table 3 of the original [paper](https://arxiv.org/pdf/2006.11477.pdf).
+Fine-tuning of `wav2vec2-large-lv60` on 100h of Librispeech training data. Results are a bit worse than those reported in the Appendix in Table 3 of the original [paper](https://arxiv.org/pdf/2006.11477.pdf).
 
 Model was trained on *librispeech-clean-train.100* with following hyper-parameters:
 
@@ -14,6 +14,6 @@ Check: https://wandb.ai/patrickvonplaten/huggingface/reports/Project-Dashboard--
 
 *Result (WER)* on Librispeech test:
 
-| "clean" (% rel difference to results in paper) | "other" (% rel difference to results in paper) |
+| "clean" | "other" |
 |---|---|
-| To evaluate | To evaluate |
+| 4.0 | 10.3 |