---
pipeline_tag: text-to-speech
tags:
  - mlx
  - tts
  - maya1
  - apple-silicon
  - snac
base_model: maya-research/maya1
library_name: mlx
---

# maya1-mlx-bf16

MLX bfloat16 conversion of [maya-research/maya1](https://huggingface.co/maya-research/maya1) for text-to-speech on Apple Silicon. Converted using [mlx-lm](https://github.com/ml-explore/mlx-examples/tree/main/llms/mlx_lm).

> **Note:** This is the full-precision (bfloat16) variant. It offers the highest quality but runs below real-time on most hardware. Consider the [8-bit](https://huggingface.co/hashb/maya1-mlx-8bit) or [4-bit](https://huggingface.co/hashb/maya1-mlx-4bit) variants for faster inference.

## Benchmarks (M4 Max, 36GB)

| Variant | Size | Tokens/s | Real-time factor |
|---------|------|----------|-----------------|
| **bf16** | **6.2 GB** | **~51 tok/s** | **0.50x** |
| 8-bit | 3.3 GB | ~91 tok/s | 0.82x |
| 4-bit | 1.8 GB | ~108 tok/s | 1.58x |

## Quick Start

Requires macOS with Apple Silicon and [uv](https://docs.astral.sh/uv/).

```bash
# From a text file
uv run tts.py input.txt -o output.wav

# From stdin
echo "Hello world" | uv run tts.py - -o hello.wav

# With a custom voice description
uv run tts.py input.txt -o output.wav -d "Deep male voice, British accent, slow pacing."
```

## CLI Options

| Option | Default | Description |
|--------|---------|-------------|
| `input` | *(required)* | Input text file path, or `-` for stdin |
| `-o`, `--output` | `output.wav` | Output WAV file path |
| `-d`, `--description` | Calm/clear female voice | Voice description prompt |
| `-m`, `--model` | `.` | Path to MLX model directory |
| `--max-chars` | `200` | Max characters per chunk |
| `--max-tokens` | `2048` | Max tokens per chunk |
| `--temperature` | `0.4` | Sampling temperature |
| `--top-p` | `0.9` | Top-p sampling |
| `--repetition-penalty` | `1.1` | Repetition penalty |

## Requirements

- macOS with Apple Silicon (M1 or later)
- [uv](https://docs.astral.sh/uv/) (dependencies are declared inline via PEP 723)