kani-tts-400m-en needs about 1.3 GB of memory at Q4_K_M (a 0.3 GB file) with an 8K context window, so it fits on a 4 GB GPU. This repository has 19 quants, from 0.0 GB to 0.7 GB (BF16). Longer context needs more: at 32K tokens the KV cache alone adds about 2.1 GB.
KaniTTS 400M English Text‑to‑Speech model
KaniTTS is a 400M parameter Text‑to‑Speech model optimized for real‑time conversational AI. It uses a two‑stage pipeline: a lightweight LLM generates compressed tokens, then a neural audio codec synthesizes waveforms at 22kHz. The model runs efficiently on modest GPU hardware.
Good at
- Real‑time speech synthesis
- Low‑latency inference
- Compact model size
Limits
- No multi‑speaker support
- Performance drops after 15 seconds of input
- Optimized primarily for English
Best for
- Chatbot voice output
- Edge device TTS
- Screen reader audio
Who it’s for. Developers building conversational AI or accessibility tools who need a lightweight, real‑time TTS engine. The model is suitable for edge or server deployment on GPUs with limited VRAM.
Quants & memory
| Quant | File | Est. memory | vs VRAM | Result | |
|---|---|---|---|---|---|
| — | 0.0 GB | 1.0 GB | Fits | Get file → | |
| IQ3_XXS | 0.2 GB | 1.2 GB | Fits | Get file → | |
| IQ3_XS | 0.2 GB | 1.2 GB | Fits | Get file → | |
| Q3_K_S | 0.2 GB | 1.2 GB | Fits | Get file → | |
| IQ3_M | 0.2 GB | 1.2 GB | Fits | Get file → | |
| Q3_K_M | 0.2 GB | 1.2 GB | Fits | Get file → | |
| IQ4_NL | 0.2 GB | 1.2 GB | Fits | Get file → | |
| IQ4_XS | 0.2 GB | 1.3 GB | Fits | Get file → | |
| Q4_K_S | 0.2 GB | 1.3 GB | Fits | Get file → | |
| Q4_1 | 0.2 GB | 1.3 GB | Fits | Get file → | |
| Q4_0 | 0.3 GB | 1.3 GB | Fits | Get file → | |
| Q4_K_M | 0.3 GB | 1.3 GB | Fits | Get file → | |
| Q5_0 | 0.3 GB | 1.3 GB | Fits | Get file → | |
| Q5_K_M | 0.3 GB | 1.3 GB | Fits | Get file → | |
| Q5_1 | 0.3 GB | 1.3 GB | Fits | Get file → | |
| Q6_K_M | 0.3 GB | 1.4 GB | Fits | Get file → | |
| Q8_0 | 0.4 GB | 1.4 GB | Fits | Get file → | |
| F16 | 0.6 GB | 1.6 GB | Fits | Get file → | |
| BF16 | 0.7 GB | 1.8 GB | Fits | Get file → |
For 8 GB VRAM, use IQ3_XXS or IQ3_XS (0.19 GB, ~1.2 GB memory). For 12 GB VRAM, Q4_K_M or Q4_1 (0.26 GB, ~1.3 GB memory) provide a good balance. For 16 GB VRAM, Q6_K_M or Q8_0 (0.33 GB, ~1.4 GB memory) give higher quality.
What GPU do you need for kani-tts-400m-en?
These figures add the file size, an fp16 KV cache at an 8K context window (worked out from this model’s own header) 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 BF16 (0.7 GB file, about 1.8 GB in use).
Which kani-tts-400m-en quant should you choose?
Q4_K_M is the usual starting point for kani-tts-400m-en: it balances file size and quality well. Here it is a 0.3 GB file.
The Q8_0 file is 0.4 GB, about 1.5 times the size of Q4_K_M. It keeps the weights closer to the original, which only matters if you have the memory to spare.
If memory is tight, the smallest file is the first one in the table at 0.0 GB, with the quality trade-offs that come with fewer bits.
New to quant names? Read GGUF quantization explained.
Specifications
Architecture from the GGUF header
Source
Frequently asked questions about kani-tts-400m-en
How much VRAM does kani-tts-400m-en need?
About 1.3 GB at Q4_K_M with an 8K context window. That is the 0.3 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 kani-tts-400m-en on an 8 GB GPU?
Yes, with BF16: the 0.7 GB file needs about 1.8 GB at 8K context. Larger quants in this repository do not fit in 8 GB.
What is the file size of kani-tts-400m-en GGUF?
It depends on the quant: file is 0.0 GB, IQ3_XXS is 0.2 GB, IQ3_XS is 0.2 GB, Q3_K_S is 0.2 GB, IQ3_M is 0.2 GB, Q3_K_M is 0.2 GB and 13 more.
What context length does kani-tts-400m-en support?
The GGUF header lists a trained context length of 128,000 tokens (125K). The KV cache grows with the context you use: about 0.5 GB at 8K and 2.1 GB at 32K.
Where can I download kani-tts-400m-en GGUF?
The files are on Hugging Face in the Mungert/kani-tts-400m-en-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 kani-tts-400m-en use?
Hugging Face lists the license as apache-2.0. 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 KaniTTS on 8 GB VRAM?
Yes, use IQ3_XXS or IQ3_XS quantizations. They require about 1.2 GB of VRAM at 8k context.
Does KaniTTS support multiple voices?
No, the base model does not support multiple speakers. Fine‑tuning is required for voice variation.
What license does KaniTTS use?
The model is released under the Apache 2.0 license.