reach-vb HF staff commited on
Commit
a087c76
•
1 Parent(s): 208cc28

Update README.md

Browse files
Files changed (1) hide show
  1. README.md +19 -9
README.md CHANGED
@@ -67,22 +67,32 @@ Try out Bark yourself!
67
  </a>
68
 
69
 
70
- ## 🤗 Transformers Usage
71
-
72
-
73
  You can run Bark locally with the 🤗 Transformers library from version 4.31.0 onwards.
74
 
75
- 1. First install the 🤗 [Transformers library](https://github.com/huggingface/transformers) from main:
76
 
77
  ```
78
- pip install git+https://github.com/huggingface/transformers.git
 
79
  ```
80
 
81
- 2. Run the following Python code to generate speech samples:
82
 
83
  ```python
84
- from transformers import AutoProcessor, AutoModel
 
 
 
 
 
85
 
 
 
 
 
 
 
 
86
 
87
  processor = AutoProcessor.from_pretrained("suno/bark-small")
88
  model = AutoModel.from_pretrained("suno/bark-small")
@@ -95,7 +105,7 @@ inputs = processor(
95
  speech_values = model.generate(**inputs, do_sample=True)
96
  ```
97
 
98
- 3. Listen to the speech samples either in an ipynb notebook:
99
 
100
  ```python
101
  from IPython.display import Audio
@@ -109,7 +119,7 @@ Or save them as a `.wav` file using a third-party library, e.g. `scipy`:
109
  ```python
110
  import scipy
111
 
112
- sampling_rate = model.generation_config.sample_rate
113
  scipy.io.wavfile.write("bark_out.wav", rate=sampling_rate, data=speech_values.cpu().numpy().squeeze())
114
  ```
115
 
 
67
  </a>
68
 
69
 
 
 
 
70
  You can run Bark locally with the 🤗 Transformers library from version 4.31.0 onwards.
71
 
72
+ 1. First install the 🤗 [Transformers library](https://github.com/huggingface/transformers) and scipy:
73
 
74
  ```
75
+ pip install --upgrade pip
76
+ pip install --upgrade transformers scipy
77
  ```
78
 
79
+ 2. Run inference via the `Text-to-Speech` (TTS) pipeline. You can infer the bark model via the TTS pipeline in just a few lines of code!
80
 
81
  ```python
82
+ from transformers import pipeline
83
+ import scipy
84
+
85
+ synthesiser = pipeline("text-to-speech", "suno/bark-small")
86
+
87
+ speech = pipe("Hello, my dog is cooler than you!", forward_params={"do_sample": True})
88
 
89
+ scipy.io.wavfile.write("bark_out.wav", rate=speech["sampling_rate"], data=speech["audio"])
90
+ ```
91
+
92
+ 3. Run inference via the Transformers modelling code. You can use the processor + generate code to convert text into a mono 24 kHz speech waveform for more fine-grained control.
93
+
94
+ ```python
95
+ from transformers import AutoProcessor, AutoModel
96
 
97
  processor = AutoProcessor.from_pretrained("suno/bark-small")
98
  model = AutoModel.from_pretrained("suno/bark-small")
 
105
  speech_values = model.generate(**inputs, do_sample=True)
106
  ```
107
 
108
+ 4. Listen to the speech samples either in an ipynb notebook:
109
 
110
  ```python
111
  from IPython.display import Audio
 
119
  ```python
120
  import scipy
121
 
122
+ sampling_rate = model.config.sample_rate
123
  scipy.io.wavfile.write("bark_out.wav", rate=sampling_rate, data=speech_values.cpu().numpy().squeeze())
124
  ```
125