seerie-v224-GGUF

<div align="center" style="background-color:#000000; padding:40px 20px; border-radius:8px;">

<img src="https://seerror-website.vercel.app/public/images/seerie%20logo.png" alt="Seerie Logo" width="140"><br/><br/>

<h1 style="color:#ffffff; font-size:56px; margin:10px 0 5px 0;">SEERIE</h1> <p style="margin:0 0 15px 0;"><b style="color:#ffffff; font-size:18px;">The machine scams. Seerie sees.</b></p> <p style="color:#cccccc;">Offline AI protection against digital fraud, built for India.</p> <p> <a href="https://opensource.org/licenses/Apache-2.0"><img src="https://img.shields.io/badge/License-Apache%202.0-green.svg" alt="License"></a> <a href="https://www.seerror.com"><img src="https://img.shields.io/badge/Website-seerror.com-black" alt="Website"></a> <a href="https://github.com/Seerror-Technologies"><img src="https://img.shields.io/badge/GitHub-Seerror--Technologies-181717?logo=github" alt="GitHub"></a> <a href="https://play.google.com/store/apps/details?id=com.seerror.seerie&hl=en"><img src="https://img.shields.io/badge/Google%20Play-Seerie-3DDC84?logo=googleplay&logoColor=white" alt="Google Play"></a> </p> </div>

Overview

Seerie is a compact, on-device language model that helps people in India recognize, respond to, and report digital fraud, including UPI scams, phishing, OTP and KYC fraud, digital arrest scams, and more. It runs fully offline, so user questions and personal data never leave the device.

This repository provides GGUF quantizations of Seerie for use with llama.cpp, Ollama, and LM Studio. The full-precision (safetensors) weights are available at seerror-technologies/seerie, and the Seerie app is available on Google Play.

Current release: v2 (September 2026). All files in this repository are built from the v2 model, trained on the v224 dataset.

Model details

Coverage

Topics

Responses are designed to direct users to the appropriate official channel, such as the 1930 National Cyber Crime Helpline or cybercrime.gov.in.

Languages

English, Hinglish, and 9 Indian languages in their native scripts: Hindi, Bengali, Tamil, Telugu, Kannada, Malayalam, Gujarati, Punjabi, and Odia. Seerie responds in the language of the question.

Example

User: Mera UPI se paise kat gaye, kya karoon?

Seerie: Sabse pehle turant apne bank ko call karke card/UPI block karayein.
Phir cybercrime.gov.in par complaint file karein ya 1930 helpline par call karein,
jitni jaldi report karenge, paise wapas milne ke chances utne zyada honge.
Apna transaction ID aur bank statement screenshot ready rakhein complaint ke liye.

Available files

Quick start

Ollama

ollama run hf.co/seerror-technologies/seerie-GGUF:Q4_K_M

llama.cpp

llama-cli -hf seerror-technologies/seerie-GGUF:Q4_K_M

Or with a local file:

llama-cli -m seerie-q4_k_m.gguf -p "Mera UPI se paise kat gaye, kya karoon?"

LM Studio

1. Search for seerror-technologies/seerie-GGUF in LM Studio, or download a .gguf file from this page.
2. Load the model and leave the system prompt empty.
3. Apply the recommended settings below.

Recommended settings

Evaluation

Seerie v2 was evaluated on an internal benchmark of 70 test cases across 14 categories, including UPI fraud, digital arrest scams, OTP handling, child safety, legal, medical and financial guidance boundaries, and privacy claims.

Overall score: 63/70 passed (90%)

<p align="center"> <img src="seerie_validation_results.png" alt="Evaluation results by category" width="500"> </p>

Per-case results are available in seerie_eval_results.json.

Training

<p align="center"> <img src="seerie_training_curves.png" alt="Training loss and learning rate curves" width="600"> </p>

Seerie was fine-tuned from Qwen3 1.7B using QLoRA for 3 epochs on 36,669 conversations, with the Unsloth framework for faster, memory-efficient training. The resulting model was merged and converted to GGUF with llama.cpp.

Intended use

Suitable for:

• Helping individuals identify likely scams and suspicious messages
• Explaining immediate next steps after fraud, such as blocking accounts and reporting
• Raising awareness of common fraud patterns and basic cyber law in India
• Offline or privacy-sensitive deployments on phones and laptops

Not suitable for:

• Replacing official complaints, police reports, or bank procedures
• Providing binding legal, financial, or medical advice
• Making automated decisions about individuals or transactions

Limitations

• Seerie is not a substitute for an official complaint. Urgent cases should always be reported to 1930 or cybercrime.gov.in.
• Knowledge is limited to patterns present in the training data. New scam types that emerged after training may not be recognized.
• As a 1.7B-parameter model, it is best suited to quick, practical guidance rather than detailed legal analysis.
• Like all language models, it can occasionally produce incorrect or incomplete information. Verify important details with official sources.

Changelog

v2 (September 2026)

• Retrained on dataset v224: 36,669 conversations, up from approximately 25,000
• Full 3-epoch training run
• Added support for 9 Indian languages in native scripts, up from English, Hindi, and Hinglish
• Expanded coverage to deepfake and AI voice cloning, matrimonial and romance scams, ransomware, and more
• Rebuilt all quantizations from the new model; added Q6_K and F16

v1

• Initial public release, trained on approximately 25,000 conversations in English, Hindi, and Hinglish
• Q4_K_M quantization

Citation

@misc{seerie2026,
  title        = {Seerie: An On-Device Language Model for Digital Fraud Protection in India},
  author       = {{Seerror Technologies}},
  year         = {2026},
  howpublished = {\url{https://huggingface.co/seerror-technologies/seerie-GGUF}}
}

Links

• Website: seerror.com
• GitHub: Seerror Technologies
• App: Seerie on Google Play
• Full-precision model: seerror-technologies/seerie
• Contact: [email protected]

License

Released under the Apache 2.0 License, consistent with the base model.

<div align="center"> <br/> Made in 🇮🇳 by <a href="https://www.seerror.com">Seerror Technologies</a> </div>