# Fine-tuning

!!! warning 
    We highly do note recoomand users to do fine-tuning on an RL trained model. Fine-tuning a model after RL can shift the model distribution, which may lead to degraded performance.

In the current version, you only need to finetune the 'LLAMA' part.

## Fine-tuning LLAMA
### 1. Prepare the dataset

```
.
├── SPK1
│   ├── 21.15-26.44.lab
│   ├── 21.15-26.44.mp3
│   ├── 27.51-29.98.lab
│   ├── 27.51-29.98.mp3
│   ├── 30.1-32.71.lab
│   └── 30.1-32.71.mp3
└── SPK2
    ├── 38.79-40.85.lab
    └── 38.79-40.85.mp3
```

You need to convert your dataset into the above format and place it under `data`. The audio file can have the extensions `.mp3`, `.wav`, or `.flac`, and the annotation file should have the extension `.lab`.

!!! info
    The `.lab` annotation file only needs to contain the transcription of the audio, with no special formatting required. For example, if `hi.mp3` says "Hello, goodbye," then the `hi.lab` file would contain a single line of text: "Hello, goodbye."

!!! warning
    It's recommended to apply loudness normalization to the dataset. You can use [fish-audio-preprocess](https://github.com/fishaudio/audio-preprocess) to do this.

    ```bash
    fap loudness-norm data-raw data --clean
    ```


### 2. Batch extraction of semantic tokens

Make sure you have downloaded the VQGAN weights. If not, run the following command:

```bash
huggingface-cli download fishaudio/openaudio-s1-mini --local-dir checkpoints/openaudio-s1-mini
```

You can then run the following command to extract semantic tokens:

```bash
python tools/vqgan/extract_vq.py data \
    --num-workers 1 --batch-size 16 \
    --config-name "modded_dac_vq" \
    --checkpoint-path "checkpoints/openaudio-s1-mini/codec.pth"
```

!!! note
    You can adjust `--num-workers` and `--batch-size` to increase extraction speed, but please make sure not to exceed your GPU memory limit.

This command will create `.npy` files in the `data` directory, as shown below:

```
.
├── SPK1
│   ├── 21.15-26.44.lab
│   ├── 21.15-26.44.mp3
│   ├── 21.15-26.44.npy
│   ├── 27.51-29.98.lab
│   ├── 27.51-29.98.mp3
│   ├── 27.51-29.98.npy
│   ├── 30.1-32.71.lab
│   ├── 30.1-32.71.mp3
│   └── 30.1-32.71.npy
└── SPK2
    ├── 38.79-40.85.lab
    ├── 38.79-40.85.mp3
    └── 38.79-40.85.npy
```

### 3. Pack the dataset into protobuf

```bash
python tools/llama/build_dataset.py \
    --input "data" \
    --output "data/protos" \
    --text-extension .lab \
    --num-workers 16
```

After the command finishes executing, you should see the `protos` file in the `data` directory.

### 4. Finally, fine-tuning with LoRA

Similarly, make sure you have downloaded the `LLAMA` weights. If not, run the following command:

```bash
huggingface-cli download fishaudio/openaudio-s1-mini --local-dir checkpoints/openaudio-s1-mini
```

Finally, you can start the fine-tuning by running the following command:

```bash
python fish_speech/train.py --config-name text2semantic_finetune \
    project=$project \
    +lora@model.model.lora_config=r_8_alpha_16
```

!!! note
    You can modify the training parameters such as `batch_size`, `gradient_accumulation_steps`, etc. to fit your GPU memory by modifying `fish_speech/configs/text2semantic_finetune.yaml`.

!!! note
    For Windows users, you can use `trainer.strategy.process_group_backend=gloo` to avoid `nccl` issues.

After training is complete, you can refer to the [inference](inference.md) section to test your model.

!!! info
    By default, the model will only learn the speaker's speech patterns and not the timbre. You still need to use prompts to ensure timbre stability.
    If you want to learn the timbre, you can increase the number of training steps, but this may lead to overfitting.

After training, you need to convert the LoRA weights to regular weights before performing inference.

```bash
python tools/llama/merge_lora.py \
	--lora-config r_8_alpha_16 \
	--base-weight checkpoints/openaudio-s1-mini \
	--lora-weight results/$project/checkpoints/step_000000010.ckpt \
	--output checkpoints/openaudio-s1-mini-yth-lora/
```
!!! note
    You may also try other checkpoints. We suggest using the earliest checkpoint that meets your requirements, as they often perform better on out-of-distribution (OOD) data.