Mods / Local Models

• VVITS-Dmitri (RU)by Oleg SkutteVITS, trained on Russian dataset. It supports only Russian and is a single-speaker model (Male). VITS: MIT License, Copyright (c) 2021 Jaehyeon Kim Dmitri: CC0-1.0901 downloadsupdated 2 y ago
• SSDXL-Turbo-GGUF (Q8_0)by Oleg SkutteSDXL-Turbo is a fast generative text-to-image model that can synthesize photorealistic images from a text prompt in a single network evaluation. License: StabilityAI Community License378 downloadsupdated 2 y ago
• FFLUX.2-klein-4B-GGUF (Q4_K)by Oleg SkutteFLUX.2 4B is a 4 billion parameter rectified flow transformer capable of generating images from text descriptions, delivering state-of-the-art quality with end-to-end inference in as low as under a second. Built for applications that require real-time image generation without sacrificing quality,...217 downloadsupdated 9 mo ago
• GGemma-3-27B-IT-QAT-GGUF (Q4_0)by Oleg SkutteThe 27B instruction-tuned version of the Gemma 3 model in GGUF format using Quantization Aware Training (QAT). The GGUF corresponds to Q4_0 quantization. Thanks to QAT, the model is able to preserve similar quality as bfloat16 while significantly reducing the memory requirements to load the model...162 downloadsupdated 10 mo ago
• JJETS-LJSpeech (en-US)by Oleg SkutteJETS, trained on the LJ Speech dataset. It supports only English and is a single-speaker model. JETS: Apache 2.0 @ imdanboy CMUdict: Copyright (C) 1993-2015 Carnegie Mellon University. All rights reserved.161 downloadsupdated 3 y ago
• WWhisper-small-GGML (Q5_1)by Oleg SkutteWhisper is a pre-trained model for automatic speech recognition (ASR) and speech translation. Trained on 680k hours of labelled data, Whisper models demonstrate a strong ability to generalise to many datasets and domains without the need for fine-tuning. Whisper: MIT License, Copyright (c) 2022 O...153 downloadsupdated 2 y ago
• GGemma-3-12B-IT-QAT-GGUF (Q4_0)by Oleg SkutteThe 12B instruction-tuned version of the Gemma 3 model in GGUF format using Quantization Aware Training (QAT). The GGUF corresponds to Q4_0 quantization. Thanks to QAT, the model is able to preserve similar quality as bfloat16 while significantly reducing the memory requirements to load the model...142 downloadsupdated 11 mo ago
• VVITS-LJSpeech (en-US)by Oleg SkutteVITS, trained on the LJ Speech dataset. It supports only English and is a single-speaker model. VITS: MIT License Copyright (c) 2021 Jaehyeon Kim LJ Speech: CC0-1.0125 downloadsupdated 2 y ago
• SSparkle-4B-v1-GGUF (Q4_K_M)by Oleg SkutteSparkle-4B-v1 is a compact text generation model based on Google's Gemma 3 4B Instruct. It was fine-tuned on a synthetic dataset specifically for use within DREAMIO: AI-Powered Adventures. License: Gemma @ Oleg Skutte108 downloadsupdated 12 mo ago
• FFaun-GGUF (Q8_0)by Oleg SkutteFaun is a text-to-image model assembled from Chroma-Flash and a low-step LoRA. It is the default image generation model used in DREAMIO: AI-Powered Adventures. It produces high-quality images with excellent prompt adherence in as few as 6 sampling steps. This version requires a GPU with at least...67 downloadsupdated 10 mo ago
• TTAESDXL-GGUF (Q8_0)by Oleg SkutteTAESDXL is very tiny autoencoder which uses the same "latent API" as SDXL-VAE. MIT License Copyright (c) 2023 Ollin Boer Bohan58 downloadsupdated 3 y ago
• LLCM-SSD-1B-GGUF (Q4_K)by Oleg SkutteThis checkpoint is a LCM distilled version of segmind/SSD-1B that allows to reduce the number of inference steps to only between 2 - 8 steps. License: CreativeML Open RAIL++-M @ Stability AI56 downloadsupdated 10 mo ago
• VVITS-Kokoro (JP)by Oleg SkutteVITS, trained on the Kokoro Speech dataset specifically for DREAMIO: AI-Powered Adventures. It supports only Japanese and is a single-speaker model (Male). VITS: MIT License, Copyright (c) 2021 Jaehyeon Kim Kokoro Speech: CC0-1.050 downloadsupdated 2 y ago
• TTAESD-GGUF (Q8_0)by Oleg SkutteTAESD is very tiny autoencoder which uses the same "latent API" as Stable Diffusion's VAE. MIT License Copyright (c) 2023 Ollin Boer Bohan48 downloadsupdated 2 y ago
• VVITS-Thorsten (DE)by Oleg SkutteVITS, trained on Thorsten-Voice dataset. It supports only German and is a single-speaker model (Male). VITS: MIT License , Copyright (c) 2021 Jaehyeon Kim Thorsten-Voice: CC0-1.045 downloadsupdated 2 y ago
• VVITS-Lada (UA)by Oleg SkutteVITS, trained on Ukrainian dataset. It supports only Ukrainian and is a single-speaker model (Female). VITS: MIT License, Copyright (c) 2021 Jaehyeon Kim Lada: Apache 2.0 @ Yehor Smoliakov44 downloadsupdated 2 y ago
• VVITS-Sharvard (ES)by Oleg SkutteVITS, trained on the Sharvard dataset. It supports only Spanish and is a milti-speaker model with 2 speakers (Male and Female). VITS: MIT License, Copyright (c) 2021 Jaehyeon Kim Sharvard: CC-BY-3.0 @ Vincent Aubanel, Maria Luisa García Lecumberri & Martin Cooke34 downloadsupdated 2 y ago
• VVITS-Tom (FR)by Oleg SkutteVITS, trained on French newspaper dataset. It supports only French and is a single-speaker model (Male). VITS: MIT License, Copyright (c) 2021 Jaehyeon Kim Tom: AGPL-3.0 @ Tom Darboux30 downloadsupdated 2 y ago
• VVITS-Paola (IT)by Oleg SkutteVITS, trained on Italian dataset. It supports only Italian and is a single-speaker model (Female). VITS: MIT License, Copyright (c) 2021 Jaehyeon Kim Paola: CC0-1.022 downloadsupdated 2 y ago
• VVITS-Zeroth (KR)by Oleg SkutteVITS, trained on the Zeroth-Korean dataset specifically for DREAMIO: AI-Powered Adventures. It supports only Korean and is a multi-speaker model with 105 speakers. VITS: MIT License, Copyright (c) 2021 Jaehyeon Kim Zeroth-Korean: CC-BY-4.0 @ Lucas Jo, Wonkyum Lee20 downloadsupdated 2 y ago
• VVITS-MCSpeech (PL)by Oleg SkutteVITS, trained on the MC Speech dataset. It supports only Polish and is a single-speaker model (Male). VITS: MIT License, Copyright (c) 2021 Jaehyeon Kim MC Speech: CC0-1.018 downloadsupdated 2 y ago

21 local models mods, most downloaded first. Each one has its own page here with what it needs to run and where its author publishes it; the files themselves stay with the author and are never re-hosted on this site.