Irodori-TTS-v4.1-Small needs about 1.7 GB of memory at Q6_K (a 1.0 GB file) with an 8K context window, so it fits on a 4 GB GPU. This repository has 5 quants, from 0.9 GB (Q4_K) to 1.9 GB (F16).
Japanese TTS model with improved duration prediction
Irodori‑TTS‑v4.1‑Small is a Japanese text‑to‑speech model with a separate, retrained duration predictor. It supports voice cloning, voice design, long‑reference conditioning, and emoji‑based style control. The model is released under the MIT license.
Good at
- Accurate Japanese speech synthesis
- Voice cloning with short reference clips
- Style control via captions or emojis
Limits
- Japanese only
- Requires speech.cpp 0.8.0 or later
- Short‑reference voice cloning needs ~30 s of clean audio
Best for
- Generating Japanese narration
- Creating custom voice samples
- Adding style to TTS output
Who it’s for. Developers building Japanese TTS applications or voice‑cloning tools. Researchers needing a lightweight Japanese TTS model with separate duration prediction.
Quants & memory
| Quant | File | Est. memory | vs VRAM | Result | |
|---|---|---|---|---|---|
| Q4_K | 0.9 GB | 1.5 GB | Fits | Get file → | |
| Q5_K | 0.9 GB | 1.6 GB | Fits | Get file → | |
| Q6_K | 1.0 GB | 1.7 GB | Fits | Get file → | |
| Q8_0 | 1.2 GB | 1.8 GB | Fits | Get file → | |
| F16 | 1.9 GB | 2.6 GB | Fits | Get file → |
For 8 GB VRAM, use Q4_K (0.86 GB, ~1.5 GB memory at 8k context). For 12 GB VRAM, Q5_K (0.94 GB, ~1.6 GB memory). For 16 GB VRAM, Q6_K (1.03 GB, ~1.7 GB memory). Higher‑precision F16 (1.91 GB, ~2.6 GB) requires more memory.
What GPU do you need for Irodori-TTS-v4.1-Small?
These figures add the file size, an fp16 KV cache at an 8K context window and about 0.5 GB for the runtime. They are estimates, so leave a little headroom, and expect to use less context or a smaller quant if you are right at the limit.
- 4 GB VRAM: the best quant that fits is F16 (1.9 GB file, about 2.6 GB in use).
Which Irodori-TTS-v4.1-Small quant should you choose?
Q6_K (1.0 GB) is the middle option in this repository. Very close to the original quality for most uses, at a clearly larger size than 4-bit.
If memory is tight, the smallest file is Q4_K at 0.9 GB, with the quality trade-offs that come with fewer bits. The usual sweet spot between file size and quality, and the quant most people start with.
New to quant names? Read GGUF quantization explained.
Specifications
Architecture from the GGUF header
Source
Frequently asked questions about Irodori-TTS-v4.1-Small
How much VRAM does Irodori-TTS-v4.1-Small need?
About 1.7 GB at Q6_K with an 8K context window. That is the 1.0 GB file, plus the KV cache and roughly 0.5 GB of runtime overhead. Smaller quants need less and longer context needs more.
Can I run Irodori-TTS-v4.1-Small on an 8 GB GPU?
Yes, with F16: the 1.9 GB file needs about 2.6 GB at 8K context. Larger quants in this repository do not fit in 8 GB.
What is the file size of Irodori-TTS-v4.1-Small GGUF?
It depends on the quant: Q4_K is 0.9 GB, Q5_K is 0.9 GB, Q6_K is 1.0 GB, Q8_0 is 1.2 GB, F16 is 1.9 GB.
Where can I download Irodori-TTS-v4.1-Small GGUF?
The files are on Hugging Face in the sakasegawa/Irodori-TTS-v4.1-Small-GGUF repository. The download buttons on this page open the file on Hugging Face. ggufmodels.com does not host any model files.
What license does Irodori-TTS-v4.1-Small use?
Hugging Face lists the license as mit. Licenses can differ between the quantized files and the original model, so check the repository and the base model before commercial use.
Can I run Irodori‑TTS‑v4.1‑Small on 8 GB VRAM?
Yes, the Q4_K quantized file is 0.86 GB and uses about 1.5 GB of VRAM at 8k context.
Does the model support English text?
No, it currently supports Japanese text input only.
What is the minimum reference audio length for voice cloning?
Approximately 30 seconds of clean reference speech is recommended for optimal speaker similarity.