Update WER score

README.md

@@ -23,7 +23,7 @@ model-index:
     metrics:
        - name: Test WER
          type: wer
-         value: 40.2
+         value: 23.70
 ---
 
 # Sinai Voice Arabic Speech Recognition Model
@@ -137,6 +137,7 @@ def predict(batch):
     batch["predicted"] = processor.batch_decode(predicted)
     return batch
 test_split = test_split.map(predict, batched=True, batch_size=16, remove_columns=["speech"])
+
 transformation = jiwer.Compose([
     # normalize some diacritics, remove punctuation, and replace Persian letters with Arabic ones
     jiwer.SubstituteRegexes({
@@ -148,12 +149,24 @@ transformation = jiwer.Compose([
     jiwer.SentencesToListOfWords(),
     jiwer.RemoveEmptyStrings(),
 ])
+
 metrics = jiwer.compute_measures(
     truth=[buckwalter.trans(s) for s in test_split["sentence"]],  # Buckwalter transliteration
-    hypothesis=test_split["predicted"],
+    hypothesis=[buckwalter.trans(s) for s in test_split["predicted"]],
     truth_transform=transformation,
     hypothesis_transform=transformation,
 )
 print(f"WER: {metrics['wer']:.2%}")
 ```
-**Test Result**: 40.2%
+**Test Result**: 23.70%
+
+
+## Other Arabic Voice recognition Models
+
+الكلمات لا تكفى لشكر أولئك الذين يؤمنون أن هنالك أمل, و يسعون من أجله
+
+- [elgeish/wav2vec2-large-xlsr-53-arabic](https://huggingface.co/elgeish/wav2vec2-large-xlsr-53-arabic)
+- [othrif/wav2vec2-large-xlsr-arabic](https://huggingface.co/othrif/wav2vec2-large-xlsr-arabic)
+- [anas/wav2vec2-large-xlsr-arabic](https://huggingface.co/anas/wav2vec2-large-xlsr-arabic)
+
+