SnowFlash383935/DeepSeek-R1-0528-Qwen3-8B-abliterated-mlx

Hugging Face
TEXT GENERATIONConcurrency Cost:1Model Size:8BQuant:FP8Ctx Length:32kPublished:Jun 23, 2025License:mitArchitecture:Transformer Open Weights Warm

SnowFlash383935/DeepSeek-R1-0528-Qwen3-8B-abliterated-mlx is an MLX-converted version of the DeepSeek-R1-0528-Qwen3-8B-abliterated model. This model is designed for efficient deployment and inference on Apple Silicon, leveraging the MLX framework. It provides a readily usable format for developers working within the Apple ecosystem, enabling local execution of a Qwen3-8B-based language model. The primary use case is local inference on MLX-compatible hardware.

Loading preview...

Overview

This model, SnowFlash383935/DeepSeek-R1-0528-Qwen3-8B-abliterated-mlx, is a version of the DeepSeek-R1-0528-Qwen3-8B-abliterated model that has been converted to the MLX format. The conversion was performed using mlx-lm version 0.25.2, making it optimized for efficient execution on Apple Silicon (Macs with M-series chips).

Key Capabilities

  • MLX Compatibility: Specifically formatted for use with the MLX framework, ensuring optimized performance on Apple hardware.
  • Local Inference: Enables running a Qwen3-8B-based language model directly on compatible local devices.
  • Ease of Use: Provides straightforward Python code examples for loading and generating text with the model using the mlx_lm library.

Good For

  • Apple Silicon Users: Ideal for developers and researchers who want to run large language models locally on their M-series Macs.
  • Offline Applications: Suitable for applications requiring local, on-device inference without relying on cloud APIs.
  • Experimentation: Facilitates rapid prototyping and experimentation with a Qwen3-8B-based model in an MLX environment.