I need to run the qwen3-vl-235b model on 2 Huawei 910B devices. Could you please tell me when the vllm-ascend is expected to be updated to support the qwen3-vl-235b model?
It seems likely that there is no official timeline or announcement for when vllm-ascend will support the Qwen3-VL-235B model on Huawei 910B devices. The most recent vllm-ascend discussions and PRs focus on Atlas 300T A2 and do not mention 910B or Qwen3-VL-235B support specifically. In fact, current support is limited, and 910A/910B are not yet available, with development for more device types planned but not scheduled yet, and users are encouraged to join development if interested in broader device support. Qwen3-VL-235B is a very large model, and even on GPU, it requires at least 8×80GB cards, so hardware and software support on Ascend is likely to lag behind GPU support for some time. No roadmap or ETA is provided in the available sources for your specific request.
Would you like more detail on the current state of vllm-ascend, or on alternative ways to run Qwen3-VL-235B?
Sources:
Hint: Mention @RunLLM in the post for followups.
see here ^