Spaces:
Runtime error
Runtime error
File size: 6,486 Bytes
0d80816 |
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 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 |
# FastSpeech2 Recipe
In this recipe, we will show how to train [FastSpeech2](https://openreview.net/forum?id=piLPYqxtWuA) using Amphion's infrastructure. FastSpeech2 is a non-autoregressive TTS architecture that utilizes feed-forward Transformer blocks.
There are four stages in total:
1. Data preparation
2. Features extraction
3. Training
4. Inference
> **NOTE:** You need to run every command of this recipe in the `Amphion` root path:
> ```bash
> cd Amphion
> ```
## 1. Data Preparation
### Dataset Download
You can use the commonly used TTS dataset to train TTS model, e.g., LJSpeech, VCTK, LibriTTS, etc. We strongly recommend you use LJSpeech to train TTS model for the first time. How to download dataset is detailed [here](../../datasets/README.md).
### Configuration
After downloading the dataset, you can set the dataset paths in `exp_config.json`. Note that you can change the `dataset` list to use your preferred datasets.
```json
"dataset": [
"LJSpeech",
],
"dataset_path": {
// TODO: Fill in your dataset path
"LJSpeech": "[LJSpeech dataset path]",
},
```
## 2. Features Extraction
### Configuration
Specify the `processed_dir` and the `log_dir` and for saving the processed data and the checkpoints in `exp_config.json`:
```json
// TODO: Fill in the output log path
"log_dir": "ckpts/tts",
"preprocess": {
// TODO: Fill in the output data path
"processed_dir": "data",
...
},
```
### Run
Run the `run.sh` as the preproces stage (set `--stage 1`):
```bash
sh egs/tts/FastSpeech2/run.sh --stage 1
```
## 3. Training
### Configuration
We provide the default hyparameters in the `exp_config.json`. They can work on single NVIDIA-24g GPU. You can adjust them based on your GPU machines.
```
"train": {
"batch_size": 16,
}
```
### Run
Run the `run.sh` as the training stage (set `--stage 2`). Specify a experimental name to run the following command. The tensorboard logs and checkpoints will be saved in `ckpts/tts/[YourExptName]`.
```bash
sh egs/tts/FastSpeech2/run.sh --stage 2 --name [YourExptName]
```
> **NOTE:** The `CUDA_VISIBLE_DEVICES` is set as `"0"` in default. You can change it when running `run.sh` by specifying such as `--gpu "0,1,2,3"`.
## 4. Inference
### Configuration
For inference, you need to specify the following configurations when running `run.sh`:
| Parameters | Description | Example |
| --------------------------------------------------- | ----------------------------------------------------------------------------------------------------------------------------------------------------------------- | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ |
| `--infer_expt_dir` | The experimental directory which contains `checkpoint` | `ckpts/tts/[YourExptName]` |
| `--infer_output_dir` | The output directory to save inferred audios. | `ckpts/tts/[YourExptName]/result` |
| `--infer_mode` | The inference mode, e.g., "`single`", "`batch`". | "`single`" to generate a clip of speech, "`batch`" to generate a batch of speech at a time. |
| `--infer_dataset` | The dataset used for inference. | For LJSpeech dataset, the inference dataset would be `LJSpeech`. |
| `--infer_testing_set` | The subset of the inference dataset used for inference, e.g., train, test, golden_test | For LJSpeech dataset, the testing set would be "`test`" split from LJSpeech at the feature extraction, or "`golden_test`" cherry-picked from test set as template testing set. |
| `--infer_text` | The text to be synthesized. | "`This is a clip of generated speech with the given text from a TTS model.`" |
### Run
For example, if you want to generate speech of all testing set split from LJSpeech, just run:
```bash
sh egs/tts/FastSpeech2/run.sh --stage 3 \
--infer_expt_dir ckpts/tts/[YourExptName] \
--infer_output_dir ckpts/tts/[YourExptName]/result \
--infer_mode "batch" \
--infer_dataset "LJSpeech" \
--infer_testing_set "test"
```
Or, if you want to generate a single clip of speech from a given text, just run:
```bash
sh egs/tts/FastSpeech2/run.sh --stage 3 \
--infer_expt_dir ckpts/tts/[YourExptName] \
--infer_output_dir ckpts/tts/[YourExptName]/result \
--infer_mode "single" \
--infer_text "This is a clip of generated speech with the given text from a TTS model."
```
We will release a pre-trained FastSpeech2 model trained on LJSpeech. So you can download the pre-trained model and generate speech following the above inference instruction.
```bibtex
@inproceedings{ren2020fastspeech,
title={FastSpeech 2: Fast and High-Quality End-to-End Text to Speech},
author={Ren, Yi and Hu, Chenxu and Tan, Xu and Qin, Tao and Zhao, Sheng and Zhao, Zhou and Liu, Tie-Yan},
booktitle={International Conference on Learning Representations},
year={2020}
}
```
|