# Audio to Audio Models: Open-Weight AI Source: https://savrn.com/models/tasks/audio-to-audio Markdown alternate of the page above; the site index is https://savrn.com/llms.txt --- SAVRN Model Hub · Models by Task # Audio to Audio Models 19 open-weight audio to audio models in the SAVRN Model Hub, with LAION eV, NVIDIA and Quran Lab publishing the most. 19Models 11Publishers 10M to 481MParameter range 5Licenses ## Most Downloaded | Model | Publisher | Parameters | License | Monthly downloads | Cheapest GPUs at 16-bit | | --- | --- | --- | --- | --- | --- | | [bigvgan_v2_22khz_80band_256x](https://savrn.com/models/bigvgan-v2-22khz-80band-256x) | NVIDIA | — | mit | 1.4M | — | | [bigvgan_v2_44khz_128band_512x](https://savrn.com/models/bigvgan-v2-44khz-128band-512x) | NVIDIA | — | mit | 754.8k | — | | [RE-USE](https://savrn.com/models/re-use) | NVIDIA | 10M | other | 12.8k | 1x MI300X, $1.85/hr | | [Mini-BS-RoFormer-V2.5-46.2M](https://savrn.com/models/mini-bs-roformer-v2-5-46-2m) | HiDolen | 46M | mit | 127 | 1x MI300X, $1.85/hr | | [spansynth-edit](https://savrn.com/models/spansynth-edit) | Sungkyun Chang | 481M | apache-2.0 | — | 1x MI300X, $1.85/hr | | [DAC-16kHz-LiteRT](https://savrn.com/models/dac-16khz-litert) | LiteRT Community (FKA TFLite) | — | mit | — | — | | [ArkEcho-RVC-M3-C-F](https://savrn.com/models/arkecho-rvc-m3-c-f) | Ethernos | — | cc-by-nc-4.0 | — | — | | [MOSS-v2-12to32-Enhancer-60M](https://savrn.com/models/moss-v2-12to32-enhancer-60m) | LAION eV | — | cc-by-nc-4.0 | — | — | | [gtcrn-coreai](https://savrn.com/models/gtcrn-coreai) | Max Farrell | — | mit | — | — | | [ghana-vc-Avatime_avn](https://savrn.com/models/ghana-vc-avatime-avn) | Ghana Open AI | — | cc-by-nc-4.0 | — | — | ## Licenses | License | Models | Commercial use | | --- | --- | --- | | [mit](https://savrn.com/models/licenses/mit) | 7 | Yes | | [cc-by-nc-4.0](https://savrn.com/models/licenses/cc-by-nc-4-0) | 5 | Not without separate permission | | other | 3 | Read the license | | [cc-by-4.0](https://savrn.com/models/licenses/cc-by-4-0) | 3 | Yes | | [apache-2.0](https://savrn.com/models/licenses/apache-2-0) | 1 | Yes | ## Who Publishes Them | Publisher | Models | | --- | --- | | [LAION eV](https://savrn.com/model-publishers/laion) | 4 | | [NVIDIA](https://savrn.com/model-publishers/nvidia) | 3 | | [Quran Lab](https://savrn.com/model-publishers/quran-lab) | 2 | | [losCrossos](https://savrn.com/model-publishers/loscrossos) | 2 | | [Ghana Open AI](https://savrn.com/model-publishers/ghanaopenai) | 2 | | [Nigeiwoguolai](https://savrn.com/model-publishers/nigeiwoguolai) | 1 | ## All 19 Models Model · Audio to audio ### [bigvgan_v2_22khz_80band_256x](https://savrn.com/models/bigvgan-v2-22khz-80band-256x) [NVIDIA](https://savrn.com/model-publishers/nvidia) [[Paper]](https://arxiv.org/abs/2206.04658) - [[Code]](https://github.com/NVIDIA/BigVGAN) - [[Showcase]](https://bigvgan-demo.github.io/) - [[Project Page]](https://research.nvidia.com/labs/adlr/projects/bigvgan/) - [[Weights]](https://huggingface.co/collections/nvidia/bigvgan-66959df3d97fd7d98d97dc9a) - [[Demo]](https://huggingface.co/spaces/nvidia/BigVGAN) - General refactor and code improvements for improved readability. - Fully fused CUDA kernel of anti-alised activation (upsampling + activation + downsampling) with inference speed benchmark. - We provide pretrained checkpoints of BigVGAN-v2 using diverse audio configurations, supporting up to 44 kHz sampling rate and 512x upsampling… Open weights mit PyTorch [View model](https://savrn.com/models/bigvgan-v2-22khz-80band-256x) Model · Audio to audio ### [bigvgan_v2_44khz_128band_512x](https://savrn.com/models/bigvgan-v2-44khz-128band-512x) [NVIDIA](https://savrn.com/model-publishers/nvidia) [[Paper]](https://arxiv.org/abs/2206.04658) - [[Code]](https://github.com/NVIDIA/BigVGAN) - [[Showcase]](https://bigvgan-demo.github.io/) - [[Project Page]](https://research.nvidia.com/labs/adlr/projects/bigvgan/) - [[Weights]](https://huggingface.co/collections/nvidia/bigvgan-66959df3d97fd7d98d97dc9a) - [[Demo]](https://huggingface.co/spaces/nvidia/BigVGAN) - General refactor and code improvements for improved readability. - Fully fused CUDA kernel of anti-alised activation (upsampling + activation + downsampling) with inference speed benchmark. - We provide pretrained checkpoints of BigVGAN-v2 using diverse audio configurations, supporting up to 44 kHz sampling rate and 512x upsampling… Open weights mit PyTorch [View model](https://savrn.com/models/bigvgan-v2-44khz-128band-512x) Model · Audio to audio ### [RE-USE](https://savrn.com/models/re-use) [NVIDIA](https://savrn.com/model-publishers/nvidia) For the commercially friendly model, please refer to NVIDIA-NemotronLabs-RE-USE and NVIDIA-NemotronLabs-RE-USE-laughter In universal speech enhancement, the goal is to restore the quality of diverse degraded speech while preserving fidelity, ensuring that all other factors remain unchanged, e.g., linguistic content, speaker identity, emotion, accent, and other paralinguistic attributes. Inspired by the distortion–perception trade-off theory, our proposed single model achieves a good balance between these two objectives and has the following desirable properties: - Robustness to diverse degradations, including additive noise, reverberation, clipping, bandwidth limitation, codec artifacts… Open weights other 10M parameters mamba-ssm [View model](https://savrn.com/models/re-use) Model · Audio to audio ### [Mini-BS-RoFormer-V2.5-46.2M](https://savrn.com/models/mini-bs-roformer-v2-5-46-2m) [HiDolen](https://savrn.com/model-publishers/hidolen) Model for the music source separation task, decomposes audio into four stems: bass, drums, other, and vocals. 针对音乐音频分离的 BS-RoFormer 模型，将音频分离为 bass、drums、other、vocals。 在 MUSDB18HQ 数据的 val 集上的性能（单位 SDR，越高越好）： 使用的 transformers 库版本为 5.18.0，PyTorch 版本为 2.14.1。为了正常运行模型还需要安装库 soundfile、einops 和 librosa。 GPU 推理： 以上代码会分离出 bass、drums、other 和 vocal 四个轨道。 Mini-BS-RoFormer-V2.5 相比于 V2 的主要改进： 2. 利用注意力池化达成各 stem 共享 mask 解码器 3. 去掉 MUSDB18HQ 数据集中的坏数据进行训练 使用 MUSDB18HQ 数据集的 train 和 test 集进行训练。 学习率 3e-4，以 batchsize=8 训练 246k 步。 对 transformer 层使用 Muon 优化器，其他网络层使用 AdamW 优化器。 - https://github.com/lucidrains/BS-RoFormer - https://arxiv.org/abs/2309.02612 (Music Source Separation with Band-Split RoPE Transformer) Open weights mit 46M parameters transformers [View model](https://savrn.com/models/mini-bs-roformer-v2-5-46-2m) Model · Audio to audio ### [ArkEcho-RVC-M3-C-F](https://savrn.com/models/arkecho-rvc-m3-c-f) [Ethernos](https://savrn.com/model-publishers/ethernos) 本仓库（Repository）所包含的所有人工神经网络（Artificial Neural Network）权重文件（.pth/.index）、训练日志及相关代码，均为计算声学（Computational Acoustics）与深度学习（Deep Learning）领域的技术研究实验产物。 这些文件本质上是高维张量（High-dimensional Tensors）的数值序列，通过随机梯度下降（SGD）与反向传播算法（Backpropagation）对公开可获取的音频数据进行统计建模（Statistical Modeling）得到。其技术形态与数字图像处理中的卷积核（Convolution Kernels）、自然语言处理中的词向量（Word Embeddings）并无本质差异。 本仓库不构成对任何第三方知识产权的故意侵犯，所有代码遵循 MIT License 开源协议，模型权重文件仅作为技术实现的副产品（By-products）存在。 本技术实验所使用的训练数据集（Training Dataset）包含以下角色的语音样本： - Mon3tr（モンスター）、Mon3tr 中文语音 等 - 声源版权归属：上海鹰角网络科技有限公司（Hypergryph Network Technology Co., Ltd.）及其关联公司 - 原始作品：《明日方舟》（Arknights） 上述角色的声音版权、肖像权、姓名权及相关知识产权均完全归属于鹰角网络及其合法授权方。本实验仅基于已公开发布的游戏内语音资源进行技术层面的信号处理（Signal Processing）与特征提取（Feature… Open weights cc-by-nc-4.0 [View model](https://savrn.com/models/arkecho-rvc-m3-c-f) Model · Audio to audio ### [bigvgan_v2_22khz_80band_256x](https://savrn.com/models/bigvgan-v2-22khz-80band-256x-2) [losCrossos](https://savrn.com/model-publishers/loscrossos) [[Paper]](https://arxiv.org/abs/2206.04658) - [[Code]](https://github.com/NVIDIA/BigVGAN) - [[Showcase]](https://bigvgan-demo.github.io/) - [[Project Page]](https://research.nvidia.com/labs/adlr/projects/bigvgan/) - [[Weights]](https://huggingface.co/collections/nvidia/bigvgan-66959df3d97fd7d98d97dc9a) - [[Demo]](https://huggingface.co/spaces/nvidia/BigVGAN) - General refactor and code improvements for improved readability. - Fully fused CUDA kernel of anti-alised activation (upsampling + activation + downsampling) with inference speed benchmark. - We provide pretrained checkpoints of BigVGAN-v2 using diverse audio configurations, supporting up to 44 kHz sampling rate and 512x upsampling… Open weights mit PyTorch [View model](https://savrn.com/models/bigvgan-v2-22khz-80band-256x-2) Model · Audio to audio ### [bigvgan_v2_44khz_128band_512x](https://savrn.com/models/bigvgan-v2-44khz-128band-512x-2) [losCrossos](https://savrn.com/model-publishers/loscrossos) [[Paper]](https://arxiv.org/abs/2206.04658) - [[Code]](https://github.com/NVIDIA/BigVGAN) - [[Showcase]](https://bigvgan-demo.github.io/) - [[Project Page]](https://research.nvidia.com/labs/adlr/projects/bigvgan/) - [[Weights]](https://huggingface.co/collections/nvidia/bigvgan-66959df3d97fd7d98d97dc9a) - [[Demo]](https://huggingface.co/spaces/nvidia/BigVGAN) - General refactor and code improvements for improved readability. - Fully fused CUDA kernel of anti-alised activation (upsampling + activation + downsampling) with inference speed benchmark. - We provide pretrained checkpoints of BigVGAN-v2 using diverse audio configurations, supporting up to 44 kHz sampling rate and 512x upsampling… Open weights mit PyTorch [View model](https://savrn.com/models/bigvgan-v2-44khz-128band-512x-2) Model · Audio to audio ### [chatterbox-s3gen-vc-grow-ce-continue](https://savrn.com/models/chatterbox-s3gen-vc-grow-ce-continue) [LAION eV](https://savrn.com/model-publishers/laion) This is a research-stage audio-to-audio voice-conversion adapter, not a standalone TTS model. It changes Chatterbox's S3Gen acoustic decoder while keeping the S3 semantic tokenizer, CAMPPlus speaker encoder and HiFT vocoder frozen. The adapter is rank 128 on 224 q/k/v/out attention projections, with the previously trained 99-score conditioning projector. Download the pinned upstream Chatterbox S3Gen and the 300-hour supervised/v1 RL parent release separately. This repository does not contain the upstream base weights or the training audio. The completed v1 continue rank-128 adapter at step 4,000 is continued without merging it into the base. The reward uses the frozen model at revision… Open weights cc-by-4.0 pytorch [View model](https://savrn.com/models/chatterbox-s3gen-vc-grow-ce-continue) Model · Audio to audio ### [chatterbox-s3gen-vc-grow-ce-remerged](https://savrn.com/models/chatterbox-s3gen-vc-grow-ce-remerged) [LAION eV](https://savrn.com/model-publishers/laion) This is a research-stage audio-to-audio voice-conversion adapter, not a standalone TTS model. It changes Chatterbox's S3Gen acoustic decoder while keeping the S3 semantic tokenizer, CAMPPlus speaker encoder and HiFT vocoder frozen. The adapter is rank 128 on 224 q/k/v/out attention projections, with the previously trained 99-score conditioning projector. Download the pinned upstream Chatterbox S3Gen and the 300-hour supervised/v1 RL parent release separately. This repository does not contain the upstream base weights or the training audio. The completed v1 mergednew step-4,000 adapter is merged into its effective base and a new rank-128 adapter is initialized for this follow-up. The reward… Open weights cc-by-4.0 pytorch [View model](https://savrn.com/models/chatterbox-s3gen-vc-grow-ce-remerged) Model · Audio to audio ### [chatterbox-s3gen-vc-grow-ce-speaker](https://savrn.com/models/chatterbox-s3gen-vc-grow-ce-speaker) [LAION eV](https://savrn.com/model-publishers/laion) This is a research-stage audio-to-audio voice-conversion adapter, not a standalone TTS model. It changes Chatterbox's S3Gen acoustic decoder while keeping the S3 semantic tokenizer, CAMPPlus speaker encoder and HiFT vocoder frozen. The adapter is rank 128 on 224 q/k/v/out attention projections, with the previously trained 99-score conditioning projector. Download the pinned upstream Chatterbox S3Gen and the 300-hour supervised/v1 RL parent release separately. This repository does not contain the upstream base weights or the training audio. The completed v1 continue rank-128 adapter at step 4,000 is continued, now with an explicit target-speaker reward. The reward uses the frozen model at… Open weights cc-by-4.0 pytorch [View model](https://savrn.com/models/chatterbox-s3gen-vc-grow-ce-speaker) Model · Audio to audio ### [DAC-16kHz-LiteRT](https://savrn.com/models/dac-16khz-litert) [LiteRT Community (FKA TFLite)](https://savrn.com/model-publishers/litert-community) Descript Audio Codec running on-device on the LiteRT CompiledModel GPU (ML Drift). The convolutional encoder/decoder run on the GPU; the RVQ runs on CPU. 43:1 compression (1 s → 12×50 codes), RTF ≈ 0.82 (faster than real-time) on Pixel 8a. - dac16khzencoderfp16.tflite (43 MB) — audio[1,1,16000] → latent[1,1024,50], GPU. - dac16khzdeconlyzsfp16.tflite (105 MB) — latent[1,1024,50] → audio, GPU. - dacrvq.bin (1.2 MB) — RVQ weights (12 codebooks) for the CPU quantizer (float32 LE). encoder 367/367 + decoder 398/398 nodes on the LiteRT GPU delegate (LITERTCL, 1 partition, no CPU fallback); warm RTF ~0.82; reconstruction corr 1.0 vs PyTorch DAC. The decoder's ConvTranspose1d are rewritten to a… Open weights mit litert [View model](https://savrn.com/models/dac-16khz-litert) Model · Audio to audio ### [ghana-vc-Asante_Twi_twi](https://savrn.com/models/ghana-vc-asante-twi-twi) [Ghana Open AI](https://savrn.com/model-publishers/ghanaopenai) Voice conversion for Asante Twi (twi). An audio-to-audio model that converts speech into the Zephyr voice while preserving the language. This checkpoint is dedicated to Asante Twi: it was fine-tuned only on Asante Twi (twi) clips from starting from the zero-shot seed-uvit-whisper-small-wavenet base. Use it to convert audio in Asante Twi. For other languages — or as a cross-lingual fallback — use ghanaopenai/ghana-vc. The library auto-selects this model when the dataset/config name matches. Open weights cc-by-nc-4.0 [View model](https://savrn.com/models/ghana-vc-asante-twi-twi) Model · Audio to audio ### [ghana-vc-Avatime_avn](https://savrn.com/models/ghana-vc-avatime-avn) [Ghana Open AI](https://savrn.com/model-publishers/ghanaopenai) Voice conversion for Avatime (avn). An audio-to-audio model that converts speech into the Zephyr voice while preserving the language. This checkpoint is dedicated to Avatime: it was fine-tuned only on Avatime (avn) clips from starting from the zero-shot seed-uvit-whisper-small-wavenet base. Use it to convert audio in Avatime. For other languages — or as a cross-lingual fallback — use ghanaopenai/ghana-vc. The library auto-selects this model when the dataset/config name matches. Open weights cc-by-nc-4.0 [View model](https://savrn.com/models/ghana-vc-avatime-avn) Model · Audio to audio ### [gtcrn-coreai](https://savrn.com/models/gtcrn-coreai) [Max Farrell](https://savrn.com/model-publishers/maxffarrell) This release mirrors maxffarrell/gtcrn-coreai at source commit 5a5b0529f89c7146fc196f602ed2635adc681374. Choose DNS3 for general speech enhancement; VCTK-DEMAND is provided for matching that training corpus and reproducing related experiments. Download the complete repository, preserving the.aimodel bundle directories: For deployment you can download only exports/, LICENSE, and THIRDPARTYNOTICES.md using hf download --include. The numerical and latency results below were measured in August 2026 on the specified beta software stack; they have not been remeasured for newer releases. A reproducible, streaming Core AI conversion of GTCRN, the 48.2K-parameter speech-enhancement network… Open weights mit [View model](https://savrn.com/models/gtcrn-coreai) Model · Audio to audio ### [mingqiannailv](https://savrn.com/models/mingqiannailv) [Nigeiwoguolai](https://savrn.com/model-publishers/nigeiwoguolai) 基于明前奶绿歌声数据集 v2（1,057 段 / 50.3 分钟，全量去混响）训练的 RVC v2 声音转换模型：48 kHz、HiFi-GAN 声码器、Titan 底模微调 200 轮（26,800 步）、RMVPE 音高提取。 - mingqiannailvv2final.pth：推理用模型权重（RVC 格式，内含生成器与配置） - mingqiannailvv2.index：音色检索索引 1. 两个文件放入 logs/mingqiannailv/ 目录（没有就新建） 3. 参数建议：音高调整 0、索引率 0.3~0.5、protect 0.33 Open weights cc-by-nc-4.0 [View model](https://savrn.com/models/mingqiannailv) Model · Audio to audio ### [MOSS-v2-12to32-Enhancer-60M](https://savrn.com/models/moss-v2-12to32-enhancer-60m) [LAION eV](https://savrn.com/model-publishers/laion) A small causal Qwen3 model that predicts 32 MOSS v2 audio-codebook indices from the 12 indices of the current frame and up to ten previously generated HQ frames. It is trained from scratch on German, English, Spanish and French speech. This is an experimental enhancement model. Validation cross entropy measures token prediction under teacher forcing. It does not establish an improvement in perceived audio quality. Use the generated-history listening evaluation to judge speech content, speaker identity, artifacts and long-utterance stability. Root inference weights: step 64,699, selected by held-out extra-20-codebook cross entropy 5.818123. - Step 12,940: 100 originals / 200 input cases… Open weights cc-by-nc-4.0 pytorch [View model](https://savrn.com/models/moss-v2-12to32-enhancer-60m) Model · Audio to audio ### [sawt-v6-flash](https://savrn.com/models/sawt-v6-flash) [Quran Lab](https://savrn.com/model-publishers/quran-lab) Access requested at publisher other pytorch [View model](https://savrn.com/models/sawt-v6-flash) Model · Audio to audio ### [sawt-v6-pro](https://savrn.com/models/sawt-v6-pro) [Quran Lab](https://savrn.com/model-publishers/quran-lab) Access requested at publisher other pytorch [View model](https://savrn.com/models/sawt-v6-pro) Model · Audio to audio ### [spansynth-edit](https://savrn.com/models/spansynth-edit) [Sungkyun Chang](https://savrn.com/model-publishers/mimbres) Synthesise music from MIDI, or add, remove, and modify notes in a recording by revising its MIDI. SpanSynth-Edit generates the selected region, using surrounding audio for timbre guidance. Try it in your browser: upload audio, transcribe with YourMT3+, edit the piano roll, and generate. See the local web setup to run the app yourself. - CPU inference is also supported. Measured peak VRAM was about 3.9 GiB on a GH200 for synthesis and both editing methods with default settings (20.48 s crop, 16 steps, CFG 2.0). Install PyTorch for your GPU first, then install the CLI without downloading the demo audio: Model and codec weights download automatically on first use. No token is required.… Open weights apache-2.0 481M parameters diffusers [View model](https://savrn.com/models/spansynth-edit) ## Questions ### Which Audio to audio models are most downloaded? By monthly downloads reported by the Hugging Face Hub: bigvgan_v2_22khz_80band_256x (1.4M); bigvgan_v2_44khz_128band_512x (754.8k); RE-USE (12.8k). ## Other Tasks - [Text Generation](https://savrn.com/models/tasks/text-generation) 1548 - [Image and Text to Text](https://savrn.com/models/tasks/image-and-text-to-text) 577 - [Feature Extraction](https://savrn.com/models/tasks/feature-extraction) 398 - [Robotics](https://savrn.com/models/tasks/robotics) 398 - [Text Classification](https://savrn.com/models/tasks/text-classification) 331 - [Text to Image](https://savrn.com/models/tasks/text-to-image) 232 - [Speech Recognition](https://savrn.com/models/tasks/speech-recognition) 187 - [Text to Speech](https://savrn.com/models/tasks/text-to-speech) 145 - [Image Classification](https://savrn.com/models/tasks/image-classification) 139 - [Audio Classification](https://savrn.com/models/tasks/audio-classification) 130 - [Sentence Similarity](https://savrn.com/models/tasks/sentence-similarity) 103 - [Reinforcement Learning](https://savrn.com/models/tasks/reinforcement-learning) 100 [See all](https://savrn.com/models/tasks)