privacy filter nemotron GGUF
Overview
Highlights
- Optimized GGUF format for efficient local inference
- Specialized in high-precision PII token classification
- Apache-2.0 license for flexible commercial integration
- Reduces latency by filtering data at the edge
- Seamlessly integrates with LocalAI and llama.cpp
Usage
# Install Hugging Face transformers
pip install transformers torch
# Load model with transformers
from transformers import AutoModel, AutoTokenizer
model = AutoModel.from_pretrained("LocalAI-io/privacy-filter-nemotron-GGUF")
tokenizer = AutoTokenizer.from_pretrained("LocalAI-io/privacy-filter-nemotron-GGUF")
Hugging Face Download
We recommend downloading the model via the Hugging Face CLI or Hub SDK.
Guidance:Before downloading, install huggingface_hub with:
pip install -U huggingface_hub
CLI Download
Download the full repository
huggingface-cli download LocalAI-io/privacy-filter-nemotron-GGUF
Download a single file to a local folder (e.g. config.json into ./dir)
huggingface-cli download LocalAI-io/privacy-filter-nemotron-GGUF config.json --local-dir ./dir
See the official docs for more CLI options
SDK Download
# 模型下载
from huggingface_hub import snapshot_download
model_dir = snapshot_download('LocalAI-io/privacy-filter-nemotron-GGUF')
Git Download
Make sure git-lfs is installed first
git lfs install
git clone https://huggingface.co/LocalAI-io/privacy-filter-nemotron-GGUF
To skip LFS large-file downloads, use:
GIT_LFS_SKIP_SMUDGE=1 git clone https://huggingface.co/LocalAI-io/privacy-filter-nemotron-GGUF
Model files are hosted on the Hugging Face Hub — download directly via HF CLI / SDK / Git, not through this site.
PyTorch / Transformers Usage
Install Transformers
pip install -U transformers torch
Load the model and run inference
from transformers import AutoModelForCausalLM, AutoTokenizer
model = AutoModelForCausalLM.from_pretrained('LocalAI-io/privacy-filter-nemotron-GGUF')
tokenizer = AutoTokenizer.from_pretrained('LocalAI-io/privacy-filter-nemotron-GGUF')
Model Download
We recommend downloading the model via the ModelScope CLI or SDK.
Guidance:Before downloading, install ModelScope with:
pip install modelscope
CLI Download
Download the full repository
modelscope download --model LocalAI-io/privacy-filter-nemotron-GGUF
Download a single file to a local folder (e.g. README.md into ./dir)
modelscope download --model LocalAI-io/privacy-filter-nemotron-GGUF README.md --local_dir ./dir
See the docs for more CLI options
SDK Download
# 模型下载
from modelscope import snapshot_download
model_dir = snapshot_download('LocalAI-io/privacy-filter-nemotron-GGUF')
Git Download
Make sure git-lfs is installed first
git lfs install
git clone https://www.modelscope.cn/LocalAI-io/privacy-filter-nemotron-GGUF.git
To skip LFS large-file downloads, use:
GIT_LFS_SKIP_SMUDGE=1 git clone https://www.modelscope.cn/LocalAI-io/privacy-filter-nemotron-GGUF.git
ModelScope 模型页直接下载模型文件;无需将模型文件放在本站服务器。
Notebook Quickstart
Install the ModelScope library
pip install "modelscope[audio,cv,nlp,multi-modal,science]" -f https://modelscope.oss-cn-beijing.aliyuncs.com/releases/repo.html
Load the model and run inference
from modelscope.pipelines import pipeline
from modelscope.utils.constant import Tasks
p = pipeline('text-generation', 'LocalAI-io/privacy-filter-nemotron-GGUF')
Full Documentation
---
license: apache-2.0
base_model: OpenMed/privacy-filter-nemotron
base_model_relation: quantized
pipeline_tag: token-classification
library_name: gguf
datasets:
- nvidia/Nemotron-PII
tags:
- gguf
- privacy-filter.cpp
- llama-cpp
- localai
- token-classification
- pii
- ner
- privacy
- redaction
- nemotron
- openai-privacy-filter
language:
- en
---
privacy-filter-nemotron — GGUF (F16 + Q8_0)
GGUF conversion of OpenMed/privacy-filter-nemotron,
a fine-grained PII token-classification model — a fine-tune ofopenai/privacy-filter on thenvidia/Nemotron-PII dataset. It labels
every token with a BIOES tag over 55 PII categories (221 classes) in a single forward pass,
then decodes coherent spans with a constrained Viterbi procedure — so it can be served locally
with no Python as the encoder/NER tier of a PII redactor.
Where the base openai/privacy-filter covers 8
coarse categories, this fine-tune trades multilingual breadth for category depth: 55
fine-grained English categories (first/last name, government IDs, financial, healthcare,
vehicle, digital, …).
For the full model description, label space, evaluation, limitations, and citations, see the
source model card — this card only
covers the GGUF packaging and how to run it.
> For broader language coverage (54 categories across 16 languages) instead of this model's
> English-only depth, see the multilingual fine-tune
> privacy-filter-multilingual GGUF.
Runtimes
This GGUF uses a custom architecture, openai-privacy-filter, that is not (yet) part of
upstream llama.cpp. It runs on:
1. privacy-filter.cpp *(recommended)* —
a small standalone GGML engine for exactly this model family, on stock upstream ggml with
no patches (CPU / CUDA / Vulkan). This is the reference runtime and what the parity numbers
below are measured against.
# build (see the repo README for CUDA/Vulkan)
cmake --preset release && cmake --build --preset release -j
# run
echo "Contact John Doe at [email protected]" | \
build/release/pf-cli --classify privacy-filter-nemotron-f16.gguf 0.5It exposes a flat C API (pf_load / pf_classify → entity spans with UTF-8 byte offsets;
pf_tokenize / pf_logits) shaped for FFI — see the repo README.
2. LocalAI — install from the model gallery; LocalAI
serves it behind the gRPC TokenClassify RPC and runs the constrained BIOES Viterbi decode,
returning entity spans. LocalAI drives it through the privacy-filter backend (which
wraps privacy-filter.cpp). The model is not a chat/completion model — it is a PII detector
that other models opt into via a pii.detectors list.
3. llama.cpp — only with a patch. Stock llama.cpp, llama-cpp-python, Ollama, and
LM Studio will fail to load this file (unknown model architecture:). The arch can be added with carry-patches (TOKEN_CLS pooling, the
'openai-privacy-filter'
architecture + HF→GGUF converter, the bidirectional banded-attention graph, and an all-SWA
no-cache mask fix; TOKEN_CLS pooling tracks the still-open
PR #19725). Until that support lands
upstream, privacy-filter.cpp above is the patch-free alternative.
> Pooling note (llama.cpp path only): the model must be loaded with TOKEN_CLS pooling
> (the GGUF's default). If you drive llama-embedding directly for testing, do not pass
> --pooling none — that overrides the default and yields raw hidden states instead of label
> logits. privacy-filter.cpp handles this automatically.
Files
| File | Precision | Size | Notes |
|---|---|---|---|
| privacy-filter-nemotron-f16.gguf | F16 | 2.82 GB | Reference artifact. 156 tensors; 221 classifier.output_labels; pooling_type = TOKEN_CLS. |
| privacy-filter-nemotron-q8.gguf | Q8_0 (experts) | 1.64 GB | MoE expert weights → Q8_0, the rest F16. For RAM-constrained / edge use. |
sha256 (f16): 70dfe91ff220ff04594168a83e296dcc2054449cde77f98d0e782edbb6a31f5a
sha256 (q8): 2ec11c154e572a2686f4d77e861b7f74e6917e09638fe9bd27156d48bd99e21aQ8_0 quantization — and why it isn't free. q8 stores the bulk of the weights (the MoE
expert matrices) as 8-bit integers instead of 16-bit floats — viascripts/requant_q8.py,
with attention, embeddings and the classifier head left at F16. That cuts the download by ~42%
(2.82 GB → 1.64 GB) and is usually a bit faster on CPU.
The catch: reducing precision throws information away, and it is almost never a free lunch.
On a mixed-PII document (1,360 tokens) q8 matched f16 on 99.93% of token labels (1,359/1,360)
and produced an identical span set at threshold 0.5, with an average prediction shift (KL
divergence) of just 2.6e-5 — but note it did not match on all tokens; one token flipped.
That is the point in miniature: a reassuring average still hides the specific cases that change,
and accuracy benchmarks routinely look fine right up until the one that bites. Those numbers
also come from a single English document; a tiny *average* shift can still hide a flip on the one
input that matters to you — a rare name, an unusual ID format, or one of the fuzzier categories
below. For PII detection a single missed span is a leak, so:
- Prefer F16 if you can afford the 2.82 GB — it is the reference these numbers are measured
- Use Q8_0 when memory or speed forces it (e.g. a 4 GB Raspberry Pi 5), treat it as a
Architecture & conversion
gpt-oss-style sparse MoE (8 layers, d_model=640, 128 experts, top-4 routing; ~1.5B total /
~50M active per token), bidirectional banded attention (symmetric sliding window 128,
attention sinks retained), interleaved (GPT-J) RoPE with YaRN (θ=150000, factor 32), o200k
(o200k_base) tokenizer, and a 221-way token-classification head (score → cls.output). The
architecture is identical to the rest of the privacy-filter family — only the fine-tuned
weights and the larger (221-class) head differ.
The conversion reproduces the unmodified transformers reference at F16: across the parity
prompt set (short / PII-dense / multilingual / a 3k-token document) the F16 GGUF agrees with HF
on 99.94% of per-token argmaxes — 100% up to ~300 tokens, with the only two flips being
argmax *ties* at ~3k positions (the F16-rounding regime) — at full-logit cosine ≥ 0.9995
(mean 0.999997). A wrong expert transpose would crater that cosine, so the two load-bearing
conversion choices — the expert gate_up chunk(2) split and the n_swa = 2·sliding_window
window mapping — are confirmed by it. privacy-filter.cpp re-derives the YaRN truncate=false
frequencies at load time (fed to ggml_rope_ext as freq_factors) so the same GGUF is
interchangeable across runtimes.
This GGUF was produced by scripts/convert.py
— a self-contained HF→GGUF converter (no llama.cpp dependency). The same converter is re-run by
CI and gated against the HF reference logits for the sibling models, so the published artifact
stays in parity.
Label space
O plus B-/I-/E-/S- for each of 55 categories (1 + 55×4 = 221), spanning identity,
contact, address, dates/time, government IDs, financial, healthca