awesome-genmedia Train Rvc
Train a custom RVC (Retrieval-based Voice Conversion) model from an audio dataset. Upload a zip file containing audio splits and configure training parameters like batch size, epochs, and pitch detection method.
Quick Start
Requires an each::labs API key. Get one at eachlabs.ai.
curl -X POST https://api.eachlabs.ai/v1/prediction \
-H "Content-Type: application/json" \
-H "X-API-Key: $EACHLABS_API_KEY" \
-d '{
"model": "train-rvc",
"version": "0.0.1",
"input": {
"dataset_zip": "https://example.com/voice-dataset.zip",
"epoch": 10,
"batch_size": "7",
"sample_rate": "48k",
"f0method": "rmvpe_gpu",
"version": "v2"
}
}'
Parameters
| Parameter | Type | Default | Description |
|---|---|---|---|
| batch_size | string | 7 | Batch size for training. |
| dataset_zip | string | - | Upload dataset zip. Zip should contain dataset/<rvc_name>/split_<i>.wav. |
| epoch | integer | 10 | Number of training epochs. |
| f0method | string | rmvpe_gpu | Pitch detection method. Options: pm, dio, harvest, rmvpe, rmvpe_gpu |
| sample_rate | string | 48k | Audio sample rate. Options: 40k, 48k |
| version | string | v2 | Model version. Options: v1, v2 |
Examples
Basic training with defaults:
curl -X POST https://api.eachlabs.ai/v1/prediction \
-H "Content-Type: application/json" \
-H "X-API-Key: $EACHLABS_API_KEY" \
-d '{
"model": "train-rvc",
"version": "0.0.1",
"input": {
"dataset_zip": "https://example.com/my-voice-samples.zip",
"epoch": 15
}
}'
Extended training with specific settings:
curl -X POST https://api.eachlabs.ai/v1/prediction \
-H "Content-Type: application/json" \
-H "X-API-Key: $EACHLABS_API_KEY" \
-d '{
"model": "train-rvc",
"version": "0.0.1",
"input": {
"dataset_zip": "https://example.com/singer-dataset.zip",
"epoch": 25,
"batch_size": "12",
"sample_rate": "48k",
"f0method": "harvest",
"version": "v2"
}
}'
Related Models
- Rvc v2 - Use a trained RVC model for voice conversion
- Realistic Voice Cloning - Clone voices from audio samples