SLPL
/

SaraSadeghi commited on
Commit
b2d9405
1 Parent(s): 4a36ad9

Update README.md

Browse files
Files changed (1) hide show
  1. README.md +4 -3
README.md CHANGED
@@ -76,6 +76,8 @@ print(prediction[0])
76
  ```
77
 
78
  ## Evaluation
 
 
79
  pip install datasets
80
  pip install transformers
81
  import torch
@@ -103,6 +105,7 @@ def predict(batch):
103
  return_tensors="pt",
104
  padding=True
105
  )
 
106
  input_values = features.input_values
107
  attention_mask = features.attention_mask
108
 
@@ -117,11 +120,9 @@ dataset = dataset.map(speech_file_to_array_fn)
117
  result = dataset.map(predict, batched=True, batch_size=4)
118
  wer = load_metric("wer")
119
  cer = load_metric("cer")
 
120
  print("WER: {:.2f}".format(100 * wer.compute(predictions=result["prediction"], references=result["reference"])))
121
  print("CER: {:.2f}".format(100 * cer.compute(predictions=result["prediction"], references=result["reference"])))
122
- For the evaluation use the code below:
123
- ```python
124
- ?
125
  ```
126
 
127
  *Result (WER)*:
 
76
  ```
77
 
78
  ## Evaluation
79
+ For the evaluation use the code below:
80
+ ```python
81
  pip install datasets
82
  pip install transformers
83
  import torch
 
105
  return_tensors="pt",
106
  padding=True
107
  )
108
+
109
  input_values = features.input_values
110
  attention_mask = features.attention_mask
111
 
 
120
  result = dataset.map(predict, batched=True, batch_size=4)
121
  wer = load_metric("wer")
122
  cer = load_metric("cer")
123
+
124
  print("WER: {:.2f}".format(100 * wer.compute(predictions=result["prediction"], references=result["reference"])))
125
  print("CER: {:.2f}".format(100 * cer.compute(predictions=result["prediction"], references=result["reference"])))
 
 
 
126
  ```
127
 
128
  *Result (WER)*: