patrickvonplaten
HF staff
commited on
Commit
ca734d4
1 Parent(s): e31c31b

Update README.md

Browse files
Files changed (1) hide show
  1. README.md +5 -5
README.md CHANGED
@@ -17,7 +17,7 @@ model-index:
17
  dataset:
18
  name: arabicspeech.org MGB-3
19
  type: arabicspeech.org MGB-3
20
- args: {lang_id}
21
  metrics:
22
  - name: Test WER
23
  type: wer
@@ -40,13 +40,13 @@ resampler = torchaudio.transforms.Resample(48_000, 16_000)
40
  # Preprocessing the datasets.
41
  # We need to read the aduio files as arrays
42
  def speech_file_to_array_fn(batch):
43
- speech_array, sampling_rate = torchaudio.load(batch["path"])
44
- batch["speech"] = resampler(speech_array).squeeze().numpy()
45
- return batch
46
  test_dataset = test_dataset.map(speech_file_to_array_fn)
47
  inputs = processor(test_dataset["speech"][:2], sampling_rate=16_000, return_tensors="pt", padding=True)
48
  with torch.no_grad():
49
- logits = model(inputs.input_values, attention_mask=inputs.attention_mask).logits
50
  predicted_ids = torch.argmax(logits, dim=-1)
51
  print("Prediction:", processor.batch_decode(predicted_ids))
52
  print("Reference:", test_dataset["sentence"][:2])
17
  dataset:
18
  name: arabicspeech.org MGB-3
19
  type: arabicspeech.org MGB-3
20
+ args: ar
21
  metrics:
22
  - name: Test WER
23
  type: wer
40
  # Preprocessing the datasets.
41
  # We need to read the aduio files as arrays
42
  def speech_file_to_array_fn(batch):
43
+ \tspeech_array, sampling_rate = torchaudio.load(batch["path"])
44
+ \tbatch["speech"] = resampler(speech_array).squeeze().numpy()
45
+ \treturn batch
46
  test_dataset = test_dataset.map(speech_file_to_array_fn)
47
  inputs = processor(test_dataset["speech"][:2], sampling_rate=16_000, return_tensors="pt", padding=True)
48
  with torch.no_grad():
49
+ \tlogits = model(inputs.input_values, attention_mask=inputs.attention_mask).logits
50
  predicted_ids = torch.argmax(logits, dim=-1)
51
  print("Prediction:", processor.batch_decode(predicted_ids))
52
  print("Reference:", test_dataset["sentence"][:2])