An unofficial PyTorch implementation of [VALL-E](https://vall-e-demo.ecker.tech/), utilizing the [EnCodec](https://github.com/facebookresearch/encodec) encoder/decoder.
A script to setup a proper environment and download the weights can be invoked with `./scripts/setup.sh`. This will automatically create a `venv`, and download the `ar+nar-llama-8` weights and config file to the right place.
To quickly test if a configuration works, you can run `python -m vall_e.models.ar_nar --yaml="./data/config.yaml"`; a small trainer will overfit a provided utterance.
4. Run `python3 -m vall_e.emb.similar`. This will calculate the top-k most similar utterances for each utterance for use with sampling.
+ Doing this will help the model follow the input prompt stronger, at the possible "cost" of the model not learning how to "infer" the target speaker AND prosidy.
5. Copy `./data/config.yaml` to `./training/config.yaml`. Customize the training configuration and populate your `dataset.training` list with the values stored under `./training/dataset/list.json`.
- data is stored under `./training/data/{group}/{speaker}/{id}.{enc|dac}` as a NumPy file, where `enc` is for the EnCodec/Vocos backend, and `dac` for the Descript-Audio-Codec backend.
- it is *highly* recommended to generate metadata to speed up dataset pre-load with `python3 -m vall_e.data --yaml="./training/config.yaml" --action=metadata`
- you can convert from the standard way with the following command: `python3 -m vall_e.data --yaml="./training/config.yaml"` (metadata for dataset pre-load is generated alongside HDF5 creation)
Finetuning can be done by training the full model, or using a LoRA.
Finetuning the full model is done the same way as training a model, but be sure to have the weights in the correct spot, as if you're loading them for inferencing.
For training a LoRA, add the following block to your `config.yaml`:
And that's it. Training of the LoRA is done with the same command. Depending on the rank and alpha specified, the loss may be higher than it should, as the LoRA weights are initialized to appropriately random values. I found `rank` and `alpha` of 128 works fine.
To export your LoRA weights, run `python3 -m vall_e.export --lora --yaml="./training/config.yaml"`. You *should* be able to have the LoRA weights loaded from a training checkpoint automagically for inferencing, but export them just to be safe.
Included is a helper script to parse the training metrics. Simply invoke it with, for example: `python3 -m vall_e.plot --yaml="./training/config.yaml"`
As training under `deepspeed` and Windows is not (easily) supported, under your `config.yaml`, simply change `trainer.backend` to `local` to use the local training backend.
Creature comforts like `float16`, `amp`, and multi-GPU training *should* work under the `local` backend, but extensive testing still needs to be done to ensure it all functions.
*`transformer`: a basic attention-based transformer implementation, with attention heads + feed forwards.
*`retnet`: using [TorchScale's RetNet](https://github.com/microsoft/torchscale/blob/main/torchscale/architecture/retnet.py) implementation, a retention-based approach can be used instead.
- Its implementation for MoE can also be utilized.
* [`descript-audio-codec`](https://github.com/descriptinc/descript-audio-codec): boasts better compression and quality, but has issues with model convergence.
*`flash_attn`: uses the available `flash_attn` package (including `flash_attn==1.0.9` through a funny wrapper)
*`flash_attn_v100`: uses [ZRayZzz/flash-attention-v100](https://github.com/ZRayZzz/flash-attention-v100/)'s Flash Attention for Volta (but doesn't work currently)
[ROCm/flash-attention](https://github.com/ROCm/flash-attention) currently does not support Navi3 cards (gfx11xx), so first-class support for Flash Attention is a bit of a mess on Navi3. Using the `howiejay/navi_support` branch can get inference support, but not training support (due to some error being thrown during the backwards pass) by:
This will export the latest checkpoints, for example, under `./training/ckpt/ar+nar-retnet-8/fp32.pth`, to be loaded on any system with PyTorch, and will include additional metadata, such as the symmap used, and training stats.
*`--ar-temp`: sampling temperature to use for the AR pass. During experimentation, `0.95` provides the most consistent output, but values close to it works fine.
*`--nar-temp`: sampling temperature to use for the NAR pass. During experimentation, the lower value, the better. Set to `0` to enable greedy sampling.
*`--input-prompt-length`: the maximum duration the input prompt can be (~6 seconds is fine, longer durations lead to slower generations for "better" accuracy, as long as the model was trained against such input prompt durations)
*`--min-ar-temp`: triggers the dynamic temperature pathway, adjusting the temperature based on the confidence of the best token. Acceptable values are between `[0.0, (n)ar-temp)`.
*`--top-p`: limits the sampling pool to top sum of values that equal `P`% probability in the probability distribution.
*`--top-k`: limits the sampling pool to the top `K` values in the probability distribution.
*`--repetition-penalty`: modifies the probability of tokens if they have appeared before. In the context of audio generation, this is a very iffy parameter to use.
*`--length-penalty`: (AR only) modifies the probability of the stop token based on the current sequence length. This is ***very*** finnicky due to the AR already being well correlated with the length.
The `ar+nar-tts+stt-llama-8` model has received additional training for a speech-to-text task against EnCodec-encoded audio.
Currently, the model only transcribes back into the IPA phonemes it was trained against, as an additional model or external program is required to translate the IPA phonemes back into text.
* this does make a model that can phonemize text, and unphonemize text, more desirable in the future to replace espeak (having an additional task to handle this requires additional embeddings, output heads, and possible harm to the model as actual text is not a modality the model is trained on).
The model can be prompted in creative ways to yield some interesting behaviors:
* prompting without an input audio prompt will have the model generate a random voice at the "cost" of some unintelligible utterance at the beginning of the output response (despite doing no promptless training).
* finetunes / LoRAs can benefit from this by having input audio promptless synthesis, while opting to have an input audio prompt for guidance.
* prompting with an input text prompt being the transcription of the input audio prompt will have the response follow very closely to the input prompt (despite not doing input=output training).
* this should allow for easy transcription editing without much fuss.
This tab currently only features exploring a dataset already prepared and referenced in your `config.yaml`. You can select a registered voice, and have it randomly sample an utterance.
In the future, this *should* contain the necessary niceties to process raw audio into a dataset to train/finetune through, without needing to invoke the above commands to prepare the dataset.
So far, this only allows you to load a different model without needing to restart. The previous model should seamlessly unload, and the new one will load in place.
- something similar to HiFiGAN (or the one for TorToiSe) trained on the last hidden states of the AR *might* also enable an alternate way for streaming.
- espeak is nice, but I can only really put my whole trust with phonemizing English.
- a small model trained to handle converting text to phonemes might work, but has it's own problems (another model to carry around, as accurate as the dataset it was trained against, requires training for each language... etc).
Despite how lightweight it is in comparison to other TTS's I've meddled with, there are still some caveats, be it with the implementation or model weights:
* the audio embeddings have some quirks to having the AR's RVQ level 0 separate from the NAR's RVQ level 0 (sharing them caused some problems in testing)
* the trainer / dataloader assumes there are zero variations between a speaker's utterances, and thus it can extract the basics of a speaker's features rather than deeper features (like prosidy, tone, etc.) when performing inferences.
+ ~~however, trying to work around this would require training under `tts-c` (VALL-E continuous) mode or modifying an input prompt enough to where its quantized representation differs enough from the output response the prompt derives from.~~
+ to remedy this, training benefits from calculating the most similar utterances for each utterance, and using that as the input prompt for training.
* the trainer's default RVQ level distribution prioritizes lower RVQ levels over higher RVQ levels, as the lower levels contribute to the final waveform more; however, this leaves some minor artifacting that rises in the higher RVQ levels due to inaccuracy issues.
* speakers that aren't similar to an audiobook narrator voice has similarity issues due to the majority of training used `path`-based dataloader sampling instead of `speaker`-based (or `group`-based) dataloader sampling.
+ although LoRAs help a ton for fixing results for a single voice.
* On my test system (7900XTX), it seems inferencing quality depends on the moon phase; I don't know if it's a matter of ROCm nuances (since I've always found it to not be up to par with actual CUDA) or `bfloat16` (due to the model being trained under `float16`+AMP) being the culprit, but your mileage *will* vary depending on the system + dtype + sampler settings.
- [EnCodec](https://github.com/facebookresearch/encodec) is licensed under CC-BY-NC 4.0. If you use the code to generate audio quantization or perform decoding, it is important to adhere to the terms of their license.
- This implementation was originally based on [enhuiz/vall-e](https://github.com/enhuiz/vall-e), but has been heavily, heavily modified over time. Without it I would not have had a good basis to muck around and learn.
title={Neural Codec Language Models are Zero-Shot Text to Speech Synthesizers},
author={Wang, Chengyi and Chen, Sanyuan and Wu, Yu and Zhang, Ziqiang and Zhou, Long and Liu, Shujie and Chen, Zhuo and Liu, Yanqing and Wang, Huaming and Li, Jinyu and others},
journal={arXiv preprint arXiv:2301.02111},
year={2023}
}
```
```bibtex
@article{defossez2022highfi,
title={High Fidelity Neural Audio Compression},
author={Défossez, Alexandre and Copet, Jade and Synnaeve, Gabriel and Adi, Yossi},
