# Not able to run amd/Qwen3.8-27B-Quark-AWQ-INT4-W4A16 on rocm

**URL:** <https://discuss.vllm.ai/t/not-able-to-run-amd-qwen3-8-27b-quark-awq-int4-w4a16-on-rocm/2954>\
**Category:** General\
**Created:** [September 9, 2026, 9:56pm UTC](https://discuss.vllm.ai/t/not-able-to-run-amd-qwen3-8-27b-quark-awq-int4-w4a16-on-rocm/2954 "2026-09-09T21:56:41Z")\
**Posts on this page:** 2\
**Page:** 1

<div class="post-metadata">

**Author:** ![AnttiK](https://avatars.discourse-cdn.com/v4/letter/a/e79b87/32.png) [@AnttiK](https://discuss.vllm.ai/u/AnttiK)\
**Post date:** [September 9, 2026, 9:56pm UTC](https://discuss.vllm.ai/t/not-able-to-run-amd-qwen3-8-27b-quark-awq-int4-w4a16-on-rocm/2954/1 "2026-09-09T21:56:41Z")

</div>

how to serve:  
amd/Qwen3.8-27B-Quark-AWQ-INT4-W4A16  
on docker vllm rocm latest?

it throws me errors like  
AttributeError: ‘dict’ object has no attribute ‘endswith’  
[rank0]:[W909 21:53:54.659294217 ProcessGroupNCCL.cpp:1647] Warning: WARNING: destroy\_process\_group() was not called before program exit, which can leak resources. For more info, please see [Redirecting…](https://pytorch.org/docs/stable/distributed.html#shutdown) (function operator())

vllm 0.29

---

<div class="post-metadata">

**Author:** ![RunLLM](https://yyz2.discourse-cdn.com/flex036/user_avatar/discuss.vllm.ai/runllm/32/233_2.png) [@RunLLM](https://discuss.vllm.ai/u/RunLLM)\
**Post date:** [September 9, 2026, 9:56pm UTC](https://discuss.vllm.ai/t/not-able-to-run-amd-qwen3-8-27b-quark-awq-int4-w4a16-on-rocm/2954/2 "2026-09-09T21:56:58Z")

</div>

The error AttributeError: ‘dict’ object has no attribute ‘endswith’ when serving amd/Qwen3.8-27B-Quark-AWQ-INT4-W4A16 on vLLM ROCm Docker is a known ROCm-related issue. It occurs during model loading, specifically with Quark INT4 models, and is tracked in the vLLM issue tracker. There is currently no official fix or workaround documented in the data sources; the error is under investigation and requires further input from the vLLM and ROCm teams. You may want to follow or contribute to the discussion for updates and provide additional environment details as requested by maintainers for troubleshooting purposes. See the issue for more: [Qwen 3.8 quark int4 loading failed](https://github.com/vllm-project/vllm/issues/52454).

Would you like more detail on the troubleshooting steps or environment requirements?

Sources:

- [Qwen 3.8 quark int4 loading failed](https://github.com/vllm-project/vllm/issues/52454)

Hint: Mention @Herald in the post for followups.