license: mit
language:
- en
- bn
base_model: - deepseek-ai/DeepSeek-R1
Model name:
DeepSeek-R1-0528-Qwen3-8B-abliterated-Q5_0-GGUF
Format: GGUF
Quantization: Q5_0
Base model: DeepSeek-R1-0528 / Qwen3-8B
Converted by: rim89987
📌 Description
This is a GGUF-converted and quantized version of the DeepSeek-R1-0528 Qwen3-8B abliterated model.
The model is optimized for local and lightweight inference using llama.cpp and compatible runtimes.
The Q5_0 quantization offers a strong balance between performance, speed, and memory efficiency, making it suitable for systems with limited VRAM or free cloud environments (e.g., Colab, Kaggle).
🚀 Use Cases
General chat & assistant tasks
Reasoning and logical responses
Coding assistance (light to medium)
Local LLM experiments
GGUF-based inference pipelines
⚙️ Compatibility
llama.cpp
llama-cpp-python
Ollama (GGUF support)
Text-generation-webui (llama backend)
💾 System Requirements (Approx.)
RAM / VRAM: ~6–7 GB
Recommended for CPU or low-end GPU setups
🧾 Credits
Original model: DeepSeek / Qwen
GGUF conversion & quantization: rim89987