Hugging Face's logo

voices
/
VCTK_Canadian_English

Transformers
Model card Files
xet
Community
VCTK_Canadian_English
File size: 2,498 Bytes 
3b8d67a
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
 
import json
import os
import subprocess

# Load the data from the provided dictionary
data = {
        "VCTK_p303": {"age": 24, "gender": "F", "accents": "Canadian", "region": "Toronto", "comments": ""},
        "VCTK_p312": {"age": 19, "gender": "F", "accents": "Canadian", "region": "Hamilton", "comments": ""},
        "VCTK_p317": {"age": 23, "gender": "F", "accents": "Canadian", "region": "Hamilton", "comments": ""},
        "VCTK_p343": {"age": 27, "gender": "F", "accents": "Canadian", "region": "Alberta", "comments": ""},
        "VCTK_p307": {"age": 23, "gender": "F", "accents": "Canadian", "region": "Ontario", "comments": ""},
        "VCTK_p316": {"age": 20, "gender": "M", "accents": "Canadian", "region": "Alberta", "comments": ""},
        "VCTK_p363": {"age": 22, "gender": "M", "accents": "Canadian", "region": "Toronto", "comments": ""},
        "VCTK_p302": {"age": 20, "gender": "M", "accents": "Canadian", "region": "Montreal", "comments": ""}
    }


# Convert the data to JSON format
json_data = json.dumps(data, indent=2)

# Save the JSON data to a file
with open('speakers-log.json', 'w') as file:
    file.write(json_data)

# Run the TTS command to get the speaker indices
command = "tts --model_path checkpoint_40000.pth --config_path config.json --list_speaker_idxs | grep -vE '^(\s*\||\s*>|\s*$)'"
output = subprocess.check_output(command, shell=True, text=True)

# Parse the JSON output into a Python dictionary
speaker_indices = eval(output)

# Load the speaker IDs from speakers.json
with open('speakers-log.json', 'r') as file:
    speaker_ids = json.load(file)


for speaker_idx in speaker_indices:
    # # Remove the 'VCTK_' prefix
    speaker_id = speaker_idx
    # speaker_id = speaker_idx.replace('VCTK_', '')

    # Lookup the speaker ID in the loaded speaker IDs
    if speaker_id in speaker_ids:
        speaker_id_json = speaker_ids[speaker_id]
    else:
        continue

    # # Generate the TTS command to create the audio file
    text = f"Hello, I am from {speaker_id_json['region']}. I hope that you will select my voice for your project. Thank you."
    # # make samples directory if it doesn't exist
    if not os.path.exists("samples"):
        os.makedirs("samples")

    out_path = f"samples/{speaker_id}.wav"
    tts_command = f"tts --text \"{text}\" --model_path checkpoint_40000.pth --language_idx en --config_path config.json --speaker_idx \"{speaker_id}\" --out_path {out_path}"

    # Execute the TTS command
    os.system(tts_command)