patrickvonplaten
HF staff
commited on
Commit
d79204c
1 Parent(s): 06c667b

Update README.md

Browse files
Files changed (1) hide show
  1. README.md +4 -4
README.md CHANGED
@@ -32,14 +32,14 @@ The original model can be found under https://github.com/pytorch/fairseq/tree/ma
32
  To transcribe audio files the model can be used as a standalone acoustic model as follows:
33
 
34
  ```python
35
- from transformers import Wav2Vec2Tokenizer, Wav2Vec2ForMaskedLM
36
  from datasets import load_dataset
37
  import soundfile as sf
38
  import torch
39
 
40
  # load model and tokenizer
41
  tokenizer = Wav2Vec2Tokenizer.from_pretrained("facebook/wav2vec2-base-960h")
42
- model = Wav2Vec2ForMaskedLM.from_pretrained("facebook/wav2vec2-base-960h")
43
 
44
  # define function to read in sound file
45
  def map_to_array(batch):
@@ -68,7 +68,7 @@ To transcribe audio files the model can be used as a standalone acoustic model a
68
 
69
  ```python
70
  from datasets import load_dataset
71
- from transformers import Wav2Vec2ForMaskedLM, Wav2Vec2Tokenizer
72
  import soundfile as sf
73
  import torch
74
  from jiwer import wer
@@ -76,7 +76,7 @@ from jiwer import wer
76
 
77
  librispeech_eval = load_dataset("librispeech_asr", "clean", split="test")
78
 
79
- model = Wav2Vec2ForMaskedLM.from_pretrained("facebook/wav2vec2-base-960h").to("cuda")
80
  tokenizer = Wav2Vec2Tokenizer.from_pretrained("facebook/wav2vec2-base-960h")
81
 
82
  def map_to_array(batch):
32
  To transcribe audio files the model can be used as a standalone acoustic model as follows:
33
 
34
  ```python
35
+ from transformers import Wav2Vec2Tokenizer, Wav2Vec2ForCTC
36
  from datasets import load_dataset
37
  import soundfile as sf
38
  import torch
39
 
40
  # load model and tokenizer
41
  tokenizer = Wav2Vec2Tokenizer.from_pretrained("facebook/wav2vec2-base-960h")
42
+ model = Wav2Vec2ForCTC.from_pretrained("facebook/wav2vec2-base-960h")
43
 
44
  # define function to read in sound file
45
  def map_to_array(batch):
68
 
69
  ```python
70
  from datasets import load_dataset
71
+ from transformers import Wav2Vec2ForCTC, Wav2Vec2Tokenizer
72
  import soundfile as sf
73
  import torch
74
  from jiwer import wer
76
 
77
  librispeech_eval = load_dataset("librispeech_asr", "clean", split="test")
78
 
79
+ model = Wav2Vec2ForCTC.from_pretrained("facebook/wav2vec2-base-960h").to("cuda")
80
  tokenizer = Wav2Vec2Tokenizer.from_pretrained("facebook/wav2vec2-base-960h")
81
 
82
  def map_to_array(batch):