Svara-TTS is an open-source multilingual text-to-speech model for Indian languages, supporting 19 languages (18 Indian languages + Indian English). This model is built based on the Orpheus-style discrete audio token method, aiming to achieve clear, expressive, and low-latency speech synthesis on ordinary GPUs/CPUs.

Audio ProcessingTransformersMultiple Languages

Zonos-v0.1 is a leading open-source text-to-speech (TTS) model, trained on over 200,000 hours of multilingual speech data, comparable in expressiveness and quality to top TTS providers. It supports zero-shot voice cloning, multilingual synthesis, and fine-grained audio control.

Audio ProcessingOnnx

The GGUF quantized version of OpenAudio is a text-to-speech synthesis tool based on the FishAudio model. It supports running through simple commands and provides a convenient speech synthesis experience.

Audio ProcessingGguf

Llasa is a text-to-speech (TTS) system based on LLaMA, which extends the capabilities of the language model by integrating speech tokens, supporting Chinese and English speech generation.

Audio ProcessingSafetensorsMultiple Languages

The Orpheus Urdu speech synthesis (TTS) model is a speech synthesis tool specifically customized for Urdu. It is fine-tuned based on the Orpheus 3B model and can convert text into natural and fluent Urdu speech.

Audio ProcessingSafetensorsOther

Phi-4-multimodal-instruct is a lightweight open-source multimodal foundation model that integrates language, vision, and speech research and datasets from Phi-3.5 and 4.0 models. It supports text, image, and audio inputs to generate text outputs, with a context length of 128K tokens.

MultimodalTransformersMultiple Languages

Phi-4-multimodal-instruct is a lightweight open-source multimodal foundation model that integrates language, vision, and speech research data from Phi-3.5 and 4.0 models. It supports text, image, and audio inputs to generate text outputs, with a context length of 128K tokens.

MultimodalTransformersMultiple Languages

Zonos-v0.1 is a leading open-weight text-to-speech model trained on over 200,000 hours of multilingual speech data, delivering expressiveness and quality comparable to or even surpassing top-tier TTS service providers.

Audio ProcessingSafetensors

A Wolof text-to-speech (TTS) model fine-tuned based on the SpeechT5 architecture, using a custom tokenizer adapted to Wolof language characteristics

Audio ProcessingSafetensorsOther

Speechless is a compact open-source text-to-semantic model (1 billion parameters) designed to directly convert audio into discrete semantic tokens without relying on traditional text-to-speech (TTS) models.

Audio ProcessingSafetensorsMultiple Languages

Speechless is a compact open-source text-to-semantic model (1 billion parameters) designed to directly convert audio into discrete semantic representation tokens without relying on traditional text-to-speech (TTS) models.

Audio ProcessingSafetensorsMultiple Languages

Fish Speech 1.5 is a multilingual text-to-speech model that supports multiple languages and can be used without an access token.

Tok Pisin text-to-speech model developed by Meta, based on VITS architecture, supporting high-quality speech synthesis

Audio ProcessingTransformers

A Russian text-to-speech model trained with the NeMo toolkit, using a combination of G2P phonetic conversion, FastPitch acoustic model, and HifiGAN vocoder to support high-quality Russian speech synthesis.

Audio ProcessingNemoOther

ToxiGen is a text classification model specifically designed for detecting implicit hate speech. It is trained on a large-scale machine-generated dataset. This model can identify text content that appears harmless on the surface but contains implicit hate meanings, providing an effective technical solution for content review and social media security.

Natural Language ProcessingTransformersEnglish