update evaluation

README.md

@@ -22,16 +22,16 @@ model-index:
     metrics:
        - name: Test WER
          type: wer
-         value: 10.06
+         value: 9.04
        - name: Test CER
          type: cer
-         value: 2.57
+         value: 2.20
        - name: Test WER (+LM)
          type: wer
-         value: 7.59
+         value: 6.75
        - name: Test CER (+LM)
          type: cer
-         value: 2.11
+         value: 1.76
   - task: 
       name: Automatic Speech Recognition
       type: automatic-speech-recognition
@@ -56,7 +56,12 @@ model-index:
 
 # XLS-R-1B-ITALIAN
 
-This model is a fine-tuned version of [facebook/wav2vec2-xls-r-1b](https://huggingface.co/facebook/wav2vec2-xls-r-1b) on the MOZILLA-FOUNDATION/COMMON_VOICE_8_0 - IT dataset.
+Fine-tuned [facebook/wav2vec2-xls-r-1b](https://huggingface.co/facebook/wav2vec2-xls-r-1b) on Italian using the [Common Voice 8](https://huggingface.co/datasets/mozilla-foundation/common_voice_8_0).
+When using this model, make sure that your speech input is sampled at 16kHz.
+
+This model has been fine-tuned thanks to the GPU credits generously given by the [OVHcloud](https://www.ovhcloud.com/en/public-cloud/ai-training/) :)
+
+The script used for training can be found here: https://github.com/jonatasgrosman/wav2vec2-sprint
 
 
 ## Evaluation Commands

mozilla-foundation_common_voice_8_0_it_test_eval_results.txt

@@ -1,2 +1,2 @@
-WER: 0.07597760028374793
-CER: 0.021155387890476704
+WER: 0.0675129084844723
+CER: 0.01761353561922042

mozilla-foundation_common_voice_8_0_it_test_eval_results_greedy.txt

@@ -1,2 +1,2 @@
-WER: 0.10066230586116814
-CER: 0.025768715741336004
+WER: 0.09043783123819138
+CER: 0.022088713826821264