sumi-ja-pii needs about 1.3 GB of memory at F32 (a 0.5 GB file) with an 8K context window, so it fits on a 4 GB GPU.
Japanese PII detection model based on ModernBERT-Ja-130M
sumi-ja-pii is a 132M parameter token classifier for Japanese personal data. It extends ModernBERT-Ja-130M with a rule layer and reversible masking. The model is trained on synthetic data and supports 21 BIO classes over 10 PII types.
Good at
- Japanese PII detection
- Reversible masking for external LLMs
- Rule‑layer format handling
Limits
- No vision support
- Requires 8,192 context length
- Apache‑2.0 license only
Best for
- Internal data sanitization
- Pre‑processing for LLM pipelines
- Compliance risk reduction
Who it’s for. Developers building Japanese text processing pipelines that need PII detection and reversible masking. The model is suitable for CPU‑only environments and can be integrated with the sumi package.
Quants & memory
| Quant | File | Est. memory | vs VRAM | Result | |
|---|---|---|---|---|---|
| F32 | 0.5 GB | 1.3 GB | Fits | Get file → |
The only available GGUF quant is F32, file size 0.53 GB. It requires approximately 1.3 GB VRAM at 8,192 context. For 8 GB VRAM, F32 is safe. No other quants are provided.
What GPU do you need for sumi-ja-pii?
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 F32 (0.5 GB file, about 1.3 GB in use).
Which sumi-ja-pii quant should you choose?
F32 (0.5 GB) is the middle option in this repository. The unquantized model. It gives original quality and needs the most memory, so it is mainly for conversion or reference.
New to quant names? Read GGUF quantization explained.
Specifications
Architecture from the GGUF header
Source
Frequently asked questions about sumi-ja-pii
How much VRAM does sumi-ja-pii need?
About 1.3 GB at F32 with an 8K context window. That is the 0.5 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 sumi-ja-pii on an 8 GB GPU?
Yes, with F32: the 0.5 GB file needs about 1.3 GB at 8K context. Larger quants in this repository do not fit in 8 GB.
What is the file size of sumi-ja-pii GGUF?
It depends on the quant: F32 is 0.5 GB.
What context length does sumi-ja-pii support?
The GGUF header lists a trained context length of 8,192 tokens (8K). The KV cache grows with the context you use: about 0.3 GB at 8K and 0.3 GB at 8K.
Where can I download sumi-ja-pii GGUF?
The files are on Hugging Face in the NagaYu/sumi-ja-pii repository. The download buttons on this page open the file on Hugging Face. ggufmodels.com does not host any model files.
What license does sumi-ja-pii 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 this model on 8 GB VRAM?
Yes. The F32 quant requires about 1.3 GB VRAM at 8,192 context, leaving ample headroom.
What license applies to this model?
The model is released under the Apache‑2.0 license.
Does this model support vision tasks?
No, it is a text‑only token classifier with no vision capabilities.