Qwen3-TTS-0.6B-Base-ANE by Tiny Trash Labs LLC: Open Model

Model Card

By Tiny Trash Labs LLC, published under apache-2.0, revision 551e82a5618b.

This is a Core ML build of Qwen/Qwen3-TTS-12Hz-0.6B-Base, the voice-cloning model. It clones a voice from a few seconds of reference audio plus its transcript, and runs it on the Neural Engine of an iPhone, iPad or Mac. It needs no GPU, MLX or ONNX Runtime. It's built for Gloam Voice Studio and its radio app by Tiny Trash Labs. The Swift runtime that loads it is the open-source QwenANE target in (see Sources/QwenANE/README.md). Requires iOS 18 / macOS 15 or later (stateful Core ML models). The set is about 1.9 GB. - Base, not CustomVoice. It clones any voice rather than offering preset speakers, and both encoders are included so a new voice can be prepared on the device. - One read is one…

Read Tiny Trash Labs LLC's full model card

Qwen3-TTS 0.6B Base on the Apple Neural Engine

This is a Core ML build of Qwen/Qwen3-TTS-12Hz-0.6B-Base, the voice-cloning model. It clones a voice from a few seconds of reference audio plus its transcript, and runs it on the Neural Engine of an iPhone, iPad or Mac. It needs no GPU, MLX or ONNX Runtime.

It's built for Gloam Voice Studio and its radio app by Tiny Trash Labs. The Swift runtime that loads it is the open-source QwenANE target in TinyTrashLabs/gloam-voice-studio (see Sources/QwenANE/README.md).

What's in it

Path	What it is	Runs on
`coreml/talker0.mlmodelc`, `coreml/talker1.mlmodelc`	The talker transformer, as two stateful chunks (KV cache inside Core ML)	Neural Engine
`coreml/cp_ane.mlmodelc`	The code predictor (15 sub-codes per frame)	Neural Engine
`coreml/upF.mlmodelc`, `coreml/upMall.mlmodelc`	The vocoder upsampler	Neural Engine
`coreml/QwenSpeechEncoder.mlmodelc`	The reference audio → codec codes encoder, for on-device voice prep	Neural Engine
`coreml/QwenSpeakerEncoder.mlmodelc`	The reference audio → speaker embedding encoder, for on-device voice prep	Neural Engine
`host/`	Tokenizer, config, the text embedding and projection, the codec embeddings	CPU (Accelerate)
`vochead/`	The vocoder head (pre-transformer and quantizer)	CPU (Accelerate)

Requires iOS 18 / macOS 15 or later (stateful Core ML models). The set is about 1.9 GB.

What's different from the other Core ML ports

Base, not CustomVoice. It clones any voice rather than offering preset speakers, and both encoders are included so a new voice can be prepared on the device.
One read is one performance. QwenTalkSession / renderBreak split long text and carry each part's text and codec frames into the next part's prompt. They draw the whole read from one sampler stream and redraw a derailed take, so pace and tone don't reset at every split.
Exact sampler. It reproduces numpy's default_rng(seed), so a seeded render is reproducible and comparable to the Python reference.
Streaming. Decoded audio arrives in 0.96 s chunks, and the first chunk can be shortened for lower latency.

Speed

In our tests it renders at about real time on an iPhone 15 Pro, at roughly 5% CPU. It also renders in the background with the screen locked. Your numbers will vary by device and thermal state.

Use

import QwenANE

let engine = try QwenANEEngine(modelsDirectory: modelsURL)      // the downloaded folder
let voice  = try QwenVoicePrep.prepare(referenceWAV: wavData, transcript: "exactly what the clip says",
                                       modelsDirectory: modelsURL)
let session = QwenTalkSession(engine: engine, voice: voice, language: "en", seedText: text)
let read = try session.renderBreak(text, onPart: { part in print(part.logLine) })
// read.samples: mono Float at read.sampleRate (24 kHz)

License and attribution

The original model is © the Qwen team (Alibaba Cloud) and licensed under the Apache License 2.0. This repository is a modified redistribution under the same license (see LICENSE). Changes from the original:

converted to Core ML (fp16, stateful talker split into two chunks)
the vocoder split between Core ML (upsampler) and CPU arrays (head)
the text embedding stored quantized

The model weights were not fine-tuned.

Only clone your own voice, or a voice you have permission to use.

Identity and Version

Repository: tinytrashlabs/Qwen3-TTS-0.6B-Base-ANE
Publisher: Tiny Trash Labs LLC
Task: Text to speech
Modality: Audio
Library: coreml
Parameters: Not stated by the source
Languages: en, zh, es, fr, de, it, pt, ja
Revision: 551e82a5618b4804a51559bc508fece40d63b6a7
First published: 2026-10-05
Last updated: 2026-10-05

Files and Weights

164 files, 2.1 GB in total. The weights are 21 files totalling 1.6 GB in bin.

Weights21 files · 1.6 GB

Configuration8 files · 83.1 KB

Tokenizer1 file · 4.8 MB

Documentation2 files · 15.3 KB

Other131 files · 423.7 MB

Repository1 file · 1.5 KB

Every file

File	Type	Size	SHA-256
coreml/QwenSpeakerEncoder.mlmodelc/analytics/coremldata.bin	Weights	243 B	c18d23425aae
coreml/QwenSpeakerEncoder.mlmodelc/coremldata.bin	Weights	397 B	c8537702d99e
coreml/QwenSpeakerEncoder.mlmodelc/weights/weight.bin	Weights	35.5 MB	4e4c1c052f69
coreml/QwenSpeechEncoder.mlmodelc/analytics/coremldata.bin	Weights	243 B	f155bfe9d3af
coreml/QwenSpeechEncoder.mlmodelc/coremldata.bin	Weights	395 B	fefb6e652cc2
coreml/QwenSpeechEncoder.mlmodelc/weights/weight.bin	Weights	225.0 MB	e309e758adb2
coreml/cp_ane.mlmodelc/analytics/coremldata.bin	Weights	243 B	699ec328e73d
coreml/cp_ane.mlmodelc/coremldata.bin	Weights	325 B	bb1911b87d80
coreml/cp_ane.mlmodelc/weights/weight.bin	Weights	279.0 MB	c2fd2d5dc49c
coreml/talker0.mlmodelc/analytics/coremldata.bin	Weights	243 B	39e57acd9221
coreml/talker0.mlmodelc/coremldata.bin	Weights	1.8 KB	ac534e979074
coreml/talker0.mlmodelc/weights/weight.bin	Weights	440.5 MB	eca6f7c3b395
coreml/talker1.mlmodelc/analytics/coremldata.bin	Weights	243 B	0ad16bd6f431
coreml/talker1.mlmodelc/coremldata.bin	Weights	1.8 KB	c3035371443e
coreml/talker1.mlmodelc/weights/weight.bin	Weights	446.8 MB	1b92edd79203
coreml/upF.mlmodelc/analytics/coremldata.bin	Weights	243 B	c41ae20762d8
coreml/upF.mlmodelc/coremldata.bin	Weights	226 B	204011d94ced
coreml/upF.mlmodelc/weights/weight.bin	Weights	128.1 MB	8505c6301e1c
coreml/upMall.mlmodelc/analytics/coremldata.bin	Weights	243 B	6bb84632e90a
coreml/upMall.mlmodelc/coremldata.bin	Weights	379 B	cf31414fbea7
coreml/upMall.mlmodelc/weights/weight.bin	Weights	82.7 MB	0ea7dd920860
coreml/QwenSpeakerEncoder.mlmodelc/metadata.json	Configuration	2.6 KB	—
coreml/QwenSpeechEncoder.mlmodelc/metadata.json	Configuration	2.5 KB	—
coreml/cp_ane.mlmodelc/metadata.json	Configuration	2.2 KB	—
coreml/talker0.mlmodelc/metadata.json	Configuration	34.3 KB	—
coreml/talker1.mlmodelc/metadata.json	Configuration	35.0 KB	—
coreml/upF.mlmodelc/metadata.json	Configuration	4.1 KB	—
coreml/upMall.mlmodelc/metadata.json	Configuration	1.9 KB	—
host/config.json	Configuration	478 B	—
LICENSE	Documentation	11.4 KB	—
README.md	Documentation	4.0 KB	—
coreml/QwenSpeakerEncoder.mlmodelc/model.mil	Other	148.6 KB	—
coreml/QwenSpeechEncoder.mlmodelc/model.mil	Other	201.8 KB	—
coreml/cp_ane.mlmodelc/model.mil	Other	1.4 MB	—
coreml/talker0.mlmodelc/model.mil	Other	678.0 KB	—
coreml/talker1.mlmodelc/model.mil	Other	684.9 KB	—
coreml/upF.mlmodelc/model.mil	Other	30.4 KB	—
coreml/upMall.mlmodelc/model.mil	Other	103.1 KB	—
host/cp_codec_embedding.npy	Other	125.8 MB	79c15c37c490
host/talker_codec_embedding.npy	Other	12.6 MB	47fa9e30f98b
host/text_embedding_biases.npy	Other	9.7 MB	68af676ba2c7
host/text_embedding_q.npy	Other	155.6 MB	d69f2e5f10c9
host/text_embedding_scales.npy	Other	9.7 MB	54b57877e53e
host/text_proj_linear_fc1_b.npy	Other	8.3 KB	2ca292516e95
host/text_proj_linear_fc1_w.npy	Other	16.8 MB	dda51d812d48
host/text_proj_linear_fc2_b.npy	Other	4.2 KB	0725b7f7d031
host/text_proj_linear_fc2_w.npy	Other	8.4 MB	622562056e35
vochead/pre_conv.conv.bias.npy	Other	2.2 KB	106a8ea5121f
vochead/pre_conv.conv.weight.npy	Other	3.1 MB	60832fc60193
vochead/pre_transformer.input_proj.bias.npy	Other	1.2 KB	c5c93a078c11
vochead/pre_transformer.input_proj.weight.npy	Other	1.0 MB	e38b121e4e1e
vochead/pre_transformer.layers.0.input_layernorm.weight.npy	Other	1.2 KB	da21ab3b6253
vochead/pre_transformer.layers.0.mlp.down_proj.weight.npy	Other	1.0 MB	c1ebb28ea6b7
vochead/pre_transformer.layers.0.mlp.gate_proj.weight.npy	Other	1.0 MB	95da6c21440e
vochead/pre_transformer.layers.0.mlp.up_proj.weight.npy	Other	1.0 MB	8bb62cdd4f72
vochead/pre_transformer.layers.0.mlp_layer_scale.scale.npy	Other	1.2 KB	1613528fa079
vochead/pre_transformer.layers.0.post_attention_layernorm.weight.npy	Other	1.2 KB	0adbdd536778
vochead/pre_transformer.layers.0.self_attn.k_proj.weight.npy	Other	1.0 MB	3b42d0e019b1
vochead/pre_transformer.layers.0.self_attn.o_proj.weight.npy	Other	1.0 MB	404dc54a1d6c
vochead/pre_transformer.layers.0.self_attn.q_proj.weight.npy	Other	1.0 MB	474b5a97b401
vochead/pre_transformer.layers.0.self_attn.v_proj.weight.npy	Other	1.0 MB	fbd3957b6fa3
vochead/pre_transformer.layers.0.self_attn_layer_scale.scale.npy	Other	1.2 KB	2e2c350efa83
vochead/pre_transformer.layers.1.input_layernorm.weight.npy	Other	1.2 KB	7f1fbc2d9a08
vochead/pre_transformer.layers.1.mlp.down_proj.weight.npy	Other	1.0 MB	dec830e45766
vochead/pre_transformer.layers.1.mlp.gate_proj.weight.npy	Other	1.0 MB	9523232d6b1b
vochead/pre_transformer.layers.1.mlp.up_proj.weight.npy	Other	1.0 MB	1d0b93bb275c
vochead/pre_transformer.layers.1.mlp_layer_scale.scale.npy	Other	1.2 KB	ddefc64f7c28
vochead/pre_transformer.layers.1.post_attention_layernorm.weight.npy	Other	1.2 KB	c673ce817a04
vochead/pre_transformer.layers.1.self_attn.k_proj.weight.npy	Other	1.0 MB	3da80a9c8899
vochead/pre_transformer.layers.1.self_attn.o_proj.weight.npy	Other	1.0 MB	7aa1c46004e3
vochead/pre_transformer.layers.1.self_attn.q_proj.weight.npy	Other	1.0 MB	a12d62630d4b
vochead/pre_transformer.layers.1.self_attn.v_proj.weight.npy	Other	1.0 MB	fa840081c1af
vochead/pre_transformer.layers.1.self_attn_layer_scale.scale.npy	Other	1.2 KB	65eb4df9712e
vochead/pre_transformer.layers.2.input_layernorm.weight.npy	Other	1.2 KB	8749c229ec14
vochead/pre_transformer.layers.2.mlp.down_proj.weight.npy	Other	1.0 MB	ff45dcde1560
vochead/pre_transformer.layers.2.mlp.gate_proj.weight.npy	Other	1.0 MB	d1309b5e61f6
vochead/pre_transformer.layers.2.mlp.up_proj.weight.npy	Other	1.0 MB	2f2c3a99bac3
vochead/pre_transformer.layers.2.mlp_layer_scale.scale.npy	Other	1.2 KB	8ec16a664295
vochead/pre_transformer.layers.2.post_attention_layernorm.weight.npy	Other	1.2 KB	38a2f2837fb3
vochead/pre_transformer.layers.2.self_attn.k_proj.weight.npy	Other	1.0 MB	5b6bc9e52539
vochead/pre_transformer.layers.2.self_attn.o_proj.weight.npy	Other	1.0 MB	553f44f11cd9
vochead/pre_transformer.layers.2.self_attn.q_proj.weight.npy	Other	1.0 MB	df51c7f28690
vochead/pre_transformer.layers.2.self_attn.v_proj.weight.npy	Other	1.0 MB	46409efa3d16
vochead/pre_transformer.layers.2.self_attn_layer_scale.scale.npy	Other	1.2 KB	3a4f1b6f4ac9
vochead/pre_transformer.layers.3.input_layernorm.weight.npy	Other	1.2 KB	c468b5037609
vochead/pre_transformer.layers.3.mlp.down_proj.weight.npy	Other	1.0 MB	2f51337ab3b1
vochead/pre_transformer.layers.3.mlp.gate_proj.weight.npy	Other	1.0 MB	955f78ec8a03
vochead/pre_transformer.layers.3.mlp.up_proj.weight.npy	Other	1.0 MB	86b0b447bf35
vochead/pre_transformer.layers.3.mlp_layer_scale.scale.npy	Other	1.2 KB	a7559c86a7eb
vochead/pre_transformer.layers.3.post_attention_layernorm.weight.npy	Other	1.2 KB	0e83f4ef34bf
vochead/pre_transformer.layers.3.self_attn.k_proj.weight.npy	Other	1.0 MB	eaa31905db68
vochead/pre_transformer.layers.3.self_attn.o_proj.weight.npy	Other	1.0 MB	52ca6e7fb2a9
vochead/pre_transformer.layers.3.self_attn.q_proj.weight.npy	Other	1.0 MB	5676d5b8c86e
vochead/pre_transformer.layers.3.self_attn.v_proj.weight.npy	Other	1.0 MB	a1e67c979145
vochead/pre_transformer.layers.3.self_attn_layer_scale.scale.npy	Other	1.2 KB	4a035ab50dd8
vochead/pre_transformer.layers.4.input_layernorm.weight.npy	Other	1.2 KB	49cd118b3971
vochead/pre_transformer.layers.4.mlp.down_proj.weight.npy	Other	1.0 MB	20ebcdf29d8f
vochead/pre_transformer.layers.4.mlp.gate_proj.weight.npy	Other	1.0 MB	4c4a3d7d5372
vochead/pre_transformer.layers.4.mlp.up_proj.weight.npy	Other	1.0 MB	db97205bba9f
vochead/pre_transformer.layers.4.mlp_layer_scale.scale.npy	Other	1.2 KB	8cb462d998e5
vochead/pre_transformer.layers.4.post_attention_layernorm.weight.npy	Other	1.2 KB	bbce035aa4a7
vochead/pre_transformer.layers.4.self_attn.k_proj.weight.npy	Other	1.0 MB	ae869243ac02
vochead/pre_transformer.layers.4.self_attn.o_proj.weight.npy	Other	1.0 MB	5638f9ef99d3
vochead/pre_transformer.layers.4.self_attn.q_proj.weight.npy	Other	1.0 MB	efbd8e54bb3f
vochead/pre_transformer.layers.4.self_attn.v_proj.weight.npy	Other	1.0 MB	1f7bf876c967
vochead/pre_transformer.layers.4.self_attn_layer_scale.scale.npy	Other	1.2 KB	e916f0041702
vochead/pre_transformer.layers.5.input_layernorm.weight.npy	Other	1.2 KB	02e08d9c5b11
vochead/pre_transformer.layers.5.mlp.down_proj.weight.npy	Other	1.0 MB	b2278c9b64e7
vochead/pre_transformer.layers.5.mlp.gate_proj.weight.npy	Other	1.0 MB	d3849ea12eea
vochead/pre_transformer.layers.5.mlp.up_proj.weight.npy	Other	1.0 MB	ded4b952d9aa
vochead/pre_transformer.layers.5.mlp_layer_scale.scale.npy	Other	1.2 KB	de1523b8b90b
vochead/pre_transformer.layers.5.post_attention_layernorm.weight.npy	Other	1.2 KB	661daca1bd6c
vochead/pre_transformer.layers.5.self_attn.k_proj.weight.npy	Other	1.0 MB	8b26e2cb9aec
vochead/pre_transformer.layers.5.self_attn.o_proj.weight.npy	Other	1.0 MB	298f74e8b794
vochead/pre_transformer.layers.5.self_attn.q_proj.weight.npy	Other	1.0 MB	b054b388899d
vochead/pre_transformer.layers.5.self_attn.v_proj.weight.npy	Other	1.0 MB	ec3c54d5027d
vochead/pre_transformer.layers.5.self_attn_layer_scale.scale.npy	Other	1.2 KB	e8a6c3616597
vochead/pre_transformer.layers.6.input_layernorm.weight.npy	Other	1.2 KB	6ab04595fdc3
vochead/pre_transformer.layers.6.mlp.down_proj.weight.npy	Other	1.0 MB	43702c8b9f5e
vochead/pre_transformer.layers.6.mlp.gate_proj.weight.npy	Other	1.0 MB	116a553f5f4c
vochead/pre_transformer.layers.6.mlp.up_proj.weight.npy	Other	1.0 MB	e233f1776627
vochead/pre_transformer.layers.6.mlp_layer_scale.scale.npy	Other	1.2 KB	4af360b27e74
vochead/pre_transformer.layers.6.post_attention_layernorm.weight.npy	Other	1.2 KB	07e9d9cc06b7
vochead/pre_transformer.layers.6.self_attn.k_proj.weight.npy	Other	1.0 MB	e922d8c25c99
vochead/pre_transformer.layers.6.self_attn.o_proj.weight.npy	Other	1.0 MB	4290cd7623fb
vochead/pre_transformer.layers.6.self_attn.q_proj.weight.npy	Other	1.0 MB	5a02888f3af1
vochead/pre_transformer.layers.6.self_attn.v_proj.weight.npy	Other	1.0 MB	0990c0272091
vochead/pre_transformer.layers.6.self_attn_layer_scale.scale.npy	Other	1.2 KB	02f5d5846c98
vochead/pre_transformer.layers.7.input_layernorm.weight.npy	Other	1.2 KB	c294d991ca94
vochead/pre_transformer.layers.7.mlp.down_proj.weight.npy	Other	1.0 MB	156e7e4ea8cd
vochead/pre_transformer.layers.7.mlp.gate_proj.weight.npy	Other	1.0 MB	bffefb4ae967
vochead/pre_transformer.layers.7.mlp.up_proj.weight.npy	Other	1.0 MB	6226b9ab6d00
vochead/pre_transformer.layers.7.mlp_layer_scale.scale.npy	Other	1.2 KB	452208e50d49
vochead/pre_transformer.layers.7.post_attention_layernorm.weight.npy	Other	1.2 KB	b6fa11bb40ae
vochead/pre_transformer.layers.7.self_attn.k_proj.weight.npy	Other	1.0 MB	20bae7fdf172
vochead/pre_transformer.layers.7.self_attn.o_proj.weight.npy	Other	1.0 MB	6d3ea4ef4a17
vochead/pre_transformer.layers.7.self_attn.q_proj.weight.npy	Other	1.0 MB	49690fe20140
vochead/pre_transformer.layers.7.self_attn.v_proj.weight.npy	Other	1.0 MB	01fc685e2207
vochead/pre_transformer.layers.7.self_attn_layer_scale.scale.npy	Other	1.2 KB	d872033b1190
vochead/pre_transformer.norm.weight.npy	Other	1.2 KB	cc0d17824a17
vochead/pre_transformer.output_proj.bias.npy	Other	2.2 KB	397241d9b006
vochead/pre_transformer.output_proj.weight.npy	Other	1.0 MB	38cc5edfe86e
vochead/quantizer.rvq_first.input_proj.weight.npy	Other	262.3 KB	560c8e0a29ba
vochead/quantizer.rvq_first.output_proj.weight.npy	Other	262.3 KB	e0ec48eb24f4
vochead/quantizer.rvq_first.vq.layers.0._codebook.embed.npy	Other	1.0 MB	e78d70e402d8
vochead/quantizer.rvq_rest.input_proj.weight.npy	Other	262.3 KB	ed909f4e68ee
vochead/quantizer.rvq_rest.output_proj.weight.npy	Other	262.3 KB	e52d3d7f4ee7
vochead/quantizer.rvq_rest.vq.layers.0._codebook.embed.npy	Other	1.0 MB	26cd33af9b2c
vochead/quantizer.rvq_rest.vq.layers.1._codebook.embed.npy	Other	1.0 MB	ec5ce8e80f75
vochead/quantizer.rvq_rest.vq.layers.10._codebook.embed.npy	Other	1.0 MB	9e5427828a4f
vochead/quantizer.rvq_rest.vq.layers.11._codebook.embed.npy	Other	1.0 MB	df159e149de9
vochead/quantizer.rvq_rest.vq.layers.12._codebook.embed.npy	Other	1.0 MB	14d8879aa182
vochead/quantizer.rvq_rest.vq.layers.13._codebook.embed.npy	Other	1.0 MB	5eda154f06e1
vochead/quantizer.rvq_rest.vq.layers.14._codebook.embed.npy	Other	1.0 MB	2b741f4b8891
vochead/quantizer.rvq_rest.vq.layers.2._codebook.embed.npy	Other	1.0 MB	5040c4c3600b
vochead/quantizer.rvq_rest.vq.layers.3._codebook.embed.npy	Other	1.0 MB	95ad0925d605
vochead/quantizer.rvq_rest.vq.layers.4._codebook.embed.npy	Other	1.0 MB	37b64ebec3e3
vochead/quantizer.rvq_rest.vq.layers.5._codebook.embed.npy	Other	1.0 MB	6cd08287e534
vochead/quantizer.rvq_rest.vq.layers.6._codebook.embed.npy	Other	1.0 MB	2a9d57b654c3
vochead/quantizer.rvq_rest.vq.layers.7._codebook.embed.npy	Other	1.0 MB	e42089114d37
vochead/quantizer.rvq_rest.vq.layers.8._codebook.embed.npy	Other	1.0 MB	a26e9606c086
vochead/quantizer.rvq_rest.vq.layers.9._codebook.embed.npy	Other	1.0 MB	15b72281594f
.gitattributes	Repository	1.5 KB	—
host/tokenizer.json	Tokenizer	4.8 MB	—

License and Download

License: apache-2.0
Access: Open weights, no gate
Download size: 1.6 GB

Download from Tiny Trash Labs LLC

Released by Tiny Trash Labs LLC through its official repository on Hugging Face. Read the license.

Built From

Derived from Qwen/Qwen3-TTS-12Hz-0.6B-Base

Memory Requirements

Precision	Weights in memory
As published	1.6 GB

Weights only, from the published parameter count; the key-value cache and runtime add to this.

Questions About Qwen3-TTS-0.6B-Base-ANE

Can I use Qwen3-TTS-0.6B-Base-ANE commercially?

Yes. Qwen3-TTS-0.6B-Base-ANE is released under Apache License 2.0. The Apache License 2.0 is a permissive open-source license. It permits commercial use, modification and redistribution. It requires keeping the license and copyright notices and any NOTICE file, stating significant changes, and it includes an express patent grant from contributors.

Similar Models

Model · Text to speech

Kokoro-82M

Hexgrad

Kokoro is an open-weight TTS model with 82 million parameters. Despite its lightweight architecture, it delivers comparable quality to larger models while being significantly faster and more cost-efficient. With Apache-licensed weights, Kokoro can be deployed anywhere from production environments to personal projects. You can run this basic cell on Google Colab. Listen to samples. For more languages and details, see Advanced Usage. Under the hood, kokoro uses misaki, a G2P library at https://github.com/hexgrad/misaki Model SHA256 Hash: 496dba118d1a58f5f3db2efc88dbdc216e0483fc89fe6e47ee1f2c53f18ad1e4 Data: Kokoro was trained exclusively on permissive/non-copyrighted audio data and IPA…

Open weights apache-2.0

View model

Model · Text to speech

audio.cpp-gguf

Audio.cpp

This directory contains audio.cpp-native GGUF conversions of multiple speech models. These files are intended for use with audio.cpp. If you enjoy the project, please star audio.cpp on GitHub and this Hugging Face repository. For conversion details, supported layouts, direct-file loading, sidecar embedding, and the latest compatibility notes, see the audio.cpp GGUF guide: - https://github.com/0xShug0/audio.cpp/blob/main/docs/gguf.md!!! Converted and quantized packages are checked with automated metrics, but perceived quality can still differ for human listeners. Please validate the exact package, backend, and route to confirm the output is acceptable for your use case. The table lists the…

Open weights other audio.cpp

View model

Model · Text to speech

XTTS-v2

Coqui.ai

ⓍTTS is a Voice generation model that lets you clone voices into different languages by using just a quick 6-second audio clip. There is no need for an excessive amount of training data that spans countless hours. This is the same or similar model to what powers Coqui Studio and Coqui API. - Supports 17 languages. - Voice cloning with just a 6-second audio clip. - Emotion and style transfer by cloning. - Cross-language voice cloning. - Multi-lingual speech generation. - 24khz sampling rate. - 2 new languages; Hungarian and Korean - Architectural improvements for speaker conditioning. - Enables the use of multiple speaker references and interpolation between speakers. - Stability…

Open weights other coqui

View model

Model · Text to speech

chatterbox

Resemble AI

Chatterbox Multilingual V3 is the latest general-purpose multilingual TTS model in the Chatterbox family. It keeps the same 0.5B model size while improving speaker similarity, reducing hallucinations, and producing more natural, conversational speech across languages. V3 is designed for broad language coverage like V2, but with stronger stability and more expressive generation. It is the recommended multilingual model for users who want one voice cloning model that works across many languages. Try it in the Chatterbox Multilingual TTS V3 Space. Alongside V3, we are releasing the Single Language Pack: dedicated finetunes for priority languages where tighter quality control, stronger…

Open weights mit chatterbox

View model

Model · Text to speech

F5-TTS

Yushen CHEN

Download F5-TTS or E2 TTS and place under ckpts/ Paper: F5-TTS: A Fairytaler that Fakes Fluent and Faithful Speech with Flow Matching

Open weights cc-by-nc-4.0 f5-tts

View model

Model · Text to speech

Kokoro-82M-v1.0-ONNX

ONNX Community

Kokoro is a frontier TTS model for its size of 82 million parameters (text in/audio out). First, install the kokoro-js library from NPM using: You can then generate speech as follows: Optionally, save the audio to a file: The model is resilient to quantization, enabling efficient high-quality speech synthesis at a fraction of the original model size.

Open weights apache-2.0 transformers.js

View model