ramgpt/skywork-swe-32b-gptqmodel-4bit

Skywork-SWE-32B GPTQ 4bit (vLLM-ready)

This is a 4-bit GPTQ quantized version of Skywork/Skywork-SWE-32B, compatible with vLLM.

• Quantization: GPTQ (4-bit)
• Group size: 128
• Format: GPTQModel (custom loader)
• Dtype: float16

Usage with vLLM

vllm serve ./skywork-swe-32b-gptqmodel-4bit \
  --quantization gptq \
  --dtype half \
  --max-model-len 5900

Credits

• Base model: Skywork/Skywork-SWE-32B
• Quantized by: ramgpt