Update app.py

app.py

@@ -7,6 +7,7 @@ os.system("pip install jiwer")
 from jiwer import wer
 os.system("pip install datasets[audio]")
 from evaluate import evaluator
+import evaluate
 from datasets import load_dataset, Audio, disable_caching, set_caching_enabled
 import gradio as gr
 
@@ -16,11 +17,7 @@ disable_caching()
 huggingface_token = os.environ["huggingface_token"]
 pipe = pipeline(model="mskov/whisper-small-esc50") 
 print(pipe)
-'''
-model = WhisperModel.from_pretrained("mskov/whisper-small-miso", use_auth_token=huggingface_token)
-feature_extractor = AutoFeatureExtractor.from_pretrained("mskov/whisper-small-miso", use_auth_token=huggingface_token)
-miso_tokenizer = WhisperTokenizer.from_pretrained("mskov/whisper-small-miso", use_auth_token=huggingface_token)
-'''
+
 dataset = load_dataset("mskov/miso_test", split="test").cast_column("audio", Audio(sampling_rate=16000))
 
 print(dataset, "and at 0[audio][array] ", dataset[0]["audio"]["array"], type(dataset[0]["audio"]["array"]), "and at audio : ", dataset[0]["audio"])
@@ -37,16 +34,8 @@ iface = gr.Interface(
 )
 
 iface.launch()
-'''
-inputs = feature_extractor(dataset[0]["audio"]["array"], return_tensors="pt")
-print("inputs ::: ", inputs, "and dataset type for good measure: ", type(dataset))
-tempDataset = dataset[0]["audio"]["array"].tostring()
-tokenized_dataset = miso_tokenizer(tempDataset)  # Tokenize the dataset
 
-input_ids = features.input_ids
-attention_mask = features.attention_mask
-'''
-'''
+
 # Evaluate the model
 model.eval()
 with torch.no_grad():
@@ -63,7 +52,7 @@ wer_score = wer(labels, predicted_text)
 
 # Print or return WER score
 print(f"Word Error Rate (WER): {wer_score}")
-'''
+
 '''
 print("check check")
 print(inputs)