README.md

---
license: cc-by-nc-sa-4.0
datasets:
- mozilla-foundation/common_voice_17_0
- bond005/sberdevices_golos_10h_crowd
- bond005/sberdevices_golos_100h_farfield
- bond005/sova_rudevices
- Aniemore/resd_annotated
language:
- ru
- en
base_model:
- SWivid/F5-TTS
---
## Overview
The F5-TTS model is finetuned for Russian and English language

## License
This model is released under the Creative Commons Attribution Non Commercial Share Alike 4.0 license, which allows for free usage, modification, and distribution

## Model Information
**Base Model**: SWivid/F5-TTS  
**Training Duration:** 813k steps  
**Dataset Duration:** 100k hours

## Train charts
![alt text](loss.svg)
![alt text](lr.svg)

## Training Configuration:
```json
{
    "exp_name": "F5TTS_Base",
    "learning_rate": 1e-05,
    "batch_size_per_gpu": 5000,
    "batch_size_type": "frame",
    "max_samples": 64,
    "grad_accumulation_steps": 1,
    "max_grad_norm": 1,
    "epochs": 1,
    "num_warmup_updates": 405764,
    "save_per_updates": 811528,
    "keep_last_n_checkpoints": 5,
    "last_per_updates": 10000,
    "finetune": true,
    "file_checkpoint_train": "",
    "tokenizer_type": "char",
    "tokenizer_file": "",
    "mixed_precision": "fp16",
    "logger": "wandb",
    "bnb_optimizer": true
}
```

## Usage Instructions
Go to [base repo](https://github.com/SWivid/F5-TTS)

## To do
- Ask in community tab

# Other links
- [Github repo](https://github.com/HotDro4illa/F5-TTS)