Does it support importing the AWQ model and then exporting it in the compressed-tensor format?

[fengyang95]

[mgoin]

It would be an interesting feature to write a converted from AutoAWQ format to compressed-tensors, this is not a priority but is possible.

[brian-dellabetta]

Hi @fengyang95 , rather than integrating with AutoAWQ we are working to integrate its AWQ algorithm into our llm-compressor project, you can track this PR -- vllm-project/llm-compressor#1177.