TADA 3b-ml — MLX

MLX-converted weights for HumeAI/tada-3b-ml. Part of tada-mlx.

Built with Llama. See LICENSE for the Llama 3.2 Community License Agreement.

Usage

git clone https://github.com/gafiatulin/tada-mlx && cd tada-mlx
uv sync

# Encode a voice
uv run tada-mlx --encode-voice ref.wav --voice-text "transcript" --save-voice voice.safetensors

# Synthesize
uv run tada-mlx --text "Hello!" --prompt voice.safetensors --model gafiatulin/tada-3b-ml-mlx

Or as a library:

from tada_mlx import load_model, generate, InferenceOptions

Requires gafiatulin/tada-codec-mlx for the decoder/encoder (downloaded automatically).

Downloads last month: 15

Safetensors

Model size

4B params

Tensor type

BF16

MLX

Hardware compatibility

Quantized

Inference Providers NEW

This model isn't deployed by any Inference Provider. 🙋 Ask for provider support

Model tree for gafiatulin/tada-3b-ml-mlx

Base model

meta-llama/Llama-3.2-3B

Finetuned

HumeAI/tada-3b-ml

Finetuned

(1)

this model

Collection including gafiatulin/tada-3b-ml-mlx

Tada MLX

Collection

tada-{1b, 3b-ml} converted to MLX • 3 items • Updated Mar 20 • 1