manu02/Qwen3-0.6B-bnb-4bit-nf4-dq

Qwen3-0.6B (Quantized)

Description

This model is a 4-bit quantized version of the original Qwen/Qwen3-0.6B model, optimized for reduced memory usage while maintaining performance.

Quantization Details

• Quantization Type: 4-bit
• bnb_4bit_quant_type: nf4
• bnb_4bit_use_double_quant: True
• bnb_4bit_compute_dtype: bfloat16
• bnb_4bit_quant_storage: uint8
• Original Footprint: 1192.10 MB (BFLOAT16)
• Quantized Footprint: 842.66 MB (UINT8)
• Memory Reduction: 29.3%

Usage

from transformers import AutoModel, AutoTokenizer

model_name = "Qwen3-0.6B-bnb-4bit-nf4"
model = AutoModel.from_pretrained(
    "manu02/Qwen3-0.6B-bnb-4bit-nf4",
)
tokenizer = AutoTokenizer.from_pretrained("manu02/Qwen3-0.6B-bnb-4bit-nf4", use_fast=True)