Commit ·
f1dd1bd
1
Parent(s): 16000ae
updt example
Browse files
README.md
CHANGED
|
@@ -130,23 +130,26 @@ import torchaudio
|
|
| 130 |
|
| 131 |
from transformers import AutoModelForCTC, Wav2Vec2ProcessorWithLM
|
| 132 |
|
| 133 |
-
|
|
|
|
|
|
|
| 134 |
processor_with_lm = Wav2Vec2ProcessorWithLM.from_pretrained("bhuang/asr-wav2vec2-french")
|
|
|
|
| 135 |
|
| 136 |
wav_path = "example.wav" # path to your audio file
|
| 137 |
waveform, sample_rate = torchaudio.load(wav_path)
|
| 138 |
waveform = waveform.squeeze(axis=0) # mono
|
| 139 |
|
| 140 |
# resample
|
| 141 |
-
if sample_rate !=
|
| 142 |
-
resampler = torchaudio.transforms.Resample(sample_rate,
|
| 143 |
waveform = resampler(waveform)
|
| 144 |
|
| 145 |
# normalize
|
| 146 |
-
input_dict = processor_with_lm(waveform, sampling_rate=
|
| 147 |
|
| 148 |
with torch.inference_mode():
|
| 149 |
-
logits = model(input_dict.input_values.to(
|
| 150 |
|
| 151 |
predicted_sentence = processor_with_lm.batch_decode(logits.cpu().numpy()).text[0]
|
| 152 |
```
|
|
@@ -159,23 +162,26 @@ import torchaudio
|
|
| 159 |
|
| 160 |
from transformers import AutoModelForCTC, Wav2Vec2Processor
|
| 161 |
|
| 162 |
-
|
|
|
|
|
|
|
| 163 |
processor = Wav2Vec2Processor.from_pretrained("bhuang/asr-wav2vec2-french")
|
|
|
|
| 164 |
|
| 165 |
wav_path = "example.wav" # path to your audio file
|
| 166 |
waveform, sample_rate = torchaudio.load(wav_path)
|
| 167 |
waveform = waveform.squeeze(axis=0) # mono
|
| 168 |
|
| 169 |
# resample
|
| 170 |
-
if sample_rate !=
|
| 171 |
-
resampler = torchaudio.transforms.Resample(sample_rate,
|
| 172 |
waveform = resampler(waveform)
|
| 173 |
|
| 174 |
# normalize
|
| 175 |
-
input_dict = processor(waveform, sampling_rate=
|
| 176 |
|
| 177 |
with torch.inference_mode():
|
| 178 |
-
logits = model(input_dict.input_values.to(
|
| 179 |
|
| 180 |
# decode
|
| 181 |
predicted_ids = torch.argmax(logits, dim=-1)
|
|
|
|
| 130 |
|
| 131 |
from transformers import AutoModelForCTC, Wav2Vec2ProcessorWithLM
|
| 132 |
|
| 133 |
+
device = torch.device("cuda" if torch.cuda.is_available() else "cpu")
|
| 134 |
+
|
| 135 |
+
model = AutoModelForCTC.from_pretrained("bhuang/asr-wav2vec2-french").to(device)
|
| 136 |
processor_with_lm = Wav2Vec2ProcessorWithLM.from_pretrained("bhuang/asr-wav2vec2-french")
|
| 137 |
+
model_sample_rate = processor_with_lm.feature_extractor.sampling_rate
|
| 138 |
|
| 139 |
wav_path = "example.wav" # path to your audio file
|
| 140 |
waveform, sample_rate = torchaudio.load(wav_path)
|
| 141 |
waveform = waveform.squeeze(axis=0) # mono
|
| 142 |
|
| 143 |
# resample
|
| 144 |
+
if sample_rate != model_sample_rate:
|
| 145 |
+
resampler = torchaudio.transforms.Resample(sample_rate, model_sample_rate)
|
| 146 |
waveform = resampler(waveform)
|
| 147 |
|
| 148 |
# normalize
|
| 149 |
+
input_dict = processor_with_lm(waveform, sampling_rate=model_sample_rate, return_tensors="pt")
|
| 150 |
|
| 151 |
with torch.inference_mode():
|
| 152 |
+
logits = model(input_dict.input_values.to(device)).logits
|
| 153 |
|
| 154 |
predicted_sentence = processor_with_lm.batch_decode(logits.cpu().numpy()).text[0]
|
| 155 |
```
|
|
|
|
| 162 |
|
| 163 |
from transformers import AutoModelForCTC, Wav2Vec2Processor
|
| 164 |
|
| 165 |
+
device = torch.device("cuda" if torch.cuda.is_available() else "cpu")
|
| 166 |
+
|
| 167 |
+
model = AutoModelForCTC.from_pretrained("bhuang/asr-wav2vec2-french").to(device)
|
| 168 |
processor = Wav2Vec2Processor.from_pretrained("bhuang/asr-wav2vec2-french")
|
| 169 |
+
model_sample_rate = processor.feature_extractor.sampling_rate
|
| 170 |
|
| 171 |
wav_path = "example.wav" # path to your audio file
|
| 172 |
waveform, sample_rate = torchaudio.load(wav_path)
|
| 173 |
waveform = waveform.squeeze(axis=0) # mono
|
| 174 |
|
| 175 |
# resample
|
| 176 |
+
if sample_rate != model_sample_rate:
|
| 177 |
+
resampler = torchaudio.transforms.Resample(sample_rate, model_sample_rate)
|
| 178 |
waveform = resampler(waveform)
|
| 179 |
|
| 180 |
# normalize
|
| 181 |
+
input_dict = processor(waveform, sampling_rate=model_sample_rate, return_tensors="pt")
|
| 182 |
|
| 183 |
with torch.inference_mode():
|
| 184 |
+
logits = model(input_dict.input_values.to(device)).logits
|
| 185 |
|
| 186 |
# decode
|
| 187 |
predicted_ids = torch.argmax(logits, dim=-1)
|