Specialized Romanian Relation Extraction (Qwen3 4B). Structured JSON output. Tags: f16 (high-precision) & q4_k_m (fast).

ollama run doitmagic/qwen3-ro-rel-extract

curl http://localhost:11434/api/chat \
  -d '{
    "model": "doitmagic/qwen3-ro-rel-extract",
    "messages": [{"role": "user", "content": "Hello!"}]
  }'

from ollama import chat

response = chat(
    model='doitmagic/qwen3-ro-rel-extract',
    messages=[{'role': 'user', 'content': 'Hello!'}],
)
print(response.message.content)

import ollama from 'ollama'

const response = await ollama.chat({
  model: 'doitmagic/qwen3-ro-rel-extract',
  messages: [{role: 'user', content: 'Hello!'}],
})
console.log(response.message.content)

Models

View all →

Name

3 models

Size

Context

Input

qwen3-ro-rel-extract:latest

8.1GB · 40K context window · Text · 1 month ago

qwen3-ro-rel-extract:latest

8.1GB

40K

Text

qwen3-ro-rel-extract:q4_k_m

2.5GB · 40K context window · Text · 1 month ago

qwen3-ro-rel-extract:q4_k_m

2.5GB

40K

Text

qwen3-ro-rel-extract:f16

8.1GB · 40K context window · Text · 1 month ago

qwen3-ro-rel-extract:f16

8.1GB

40K

Text

Readme

Qwen3 RO Rel-Extract (f16 & Q4_K_M)

Specialized Romanian Relation Extraction (Qwen3 4B).

🚀 Optimized by doitmagic

Converted and published by doitmagic: - Process: High-precision f16 conversion on local infrastructure. - Integration: Tailored for Ollama.

Technical Details

Original: https://huggingface.co/andreidima/qwen3-4b-relation-extraction-multi
Format: GGUF f16 (No quality loss)
Output: Structured JSON (Subject, Relation, Object).

Example

ollama run doitmagic/qwen3-ro-rel-extract "Popescu lucrează la eMAG." Output: [{"subiect": "Popescu", "relatie": "angajat", "obiect": "eMAG"}]

Credits: Andrei Dima (Fine-tuning) / doitmagic (Conversion & Push).