running of v0.7.1 problem

[hz0ne]

base env:
torch 2.5.1+cpu
torch-npu 2.5.1rc1
torch-optimizer 0.3.0
torchaudio 2.5.1+cpu
torchmetrics 0.10.0
torchscale 0.2.0
torchtext 0.18.0+cpu
torchvision 0.20.1+cpu

CANN 8.0

install version branch:
vllm 0.7.1
vllm-ascend 0.7.1-release

when I running the vllm server, encounter AttributeError: module 'torch_npu' has no attribute 'npu_rope'

  File "/opt/conda/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1747, in _call_impl
    return forward_call(*args, **kwargs)
  File "/opt/conda/lib/python3.10/site-packages/vllm/model_executor/custom_op.py", line 23, in forward
    return self._forward_method(*args, **kwargs)
  File "/home/admin/runtime_package/vllm-ascend/vllm_ascend/ops/rotary_embedding.py", line 41, in rope_forward_oot
    torch_npu.npu_rope(
AttributeError: module 'torch_npu' has no attribute 'npu_rope'
Traceback (most recent call last):
  File "/opt/conda/bin/vllm", line 8, in <module>
    sys.exit(main())
  File "/opt/conda/lib/python3.10/site-packages/vllm/scripts.py", line 202, in main
    args.dispatch_function(args)
  File "/opt/conda/lib/python3.10/site-packages/vllm/scripts.py", line 42, in serve
    uvloop.run(run_server(args))
  File "/opt/conda/lib/python3.10/site-packages/uvloop/__init__.py", line 82, in run
    return loop.run_until_complete(wrapper())
  File "uvloop/loop.pyx", line 1518, in uvloop.loop.Loop.run_until_complete
  File "/opt/conda/lib/python3.10/site-packages/uvloop/__init__.py", line 61, in wrapper

[wangxiyuan]

0.7.1 require a new torch_npu version which is will be release in these 2-3 days And we'll release 0.7.1 official soon. Please wait more. Thanks.

[hz0ne]

0.7.1 require a new torch_npu version which is will be release in these 2-3 days And we'll release 0.7.1 official soon. Please wait more. Thanks.

OK, I tried with the main branch, and it works. The project README may need a more detailed explanation of the version branches and their corresponding features to help peoples install package.

[wangxiyuan]

yes! We're working on doc updating and will make sure the doc works before the first release.

[Yikun]

@hz0ne I added some notes for verisoing, branch and release policy, would you mind taking a look to let me know is it clear enough.

Many thanks for your feedback again.

#62

[hz0ne]

@hz0ne I added some notes for verisoing, branch and release policy, would you mind taking a look to let me know is it clear enough.

Many thanks for your feedback again.

#62

This is a relatively good version release strategy document. Thank you for your support!

[MengqingCao]

Plz try again with the torch-npu listed here:
https://vllm-ascend.readthedocs.io/en/latest/installation.html#setup-vllm-and-vllm-ascend

[wangxiyuan]

Yes, please try the new release https://github.com/vllm-project/vllm-ascend/releases/tag/v0.7.1rc1

[dawnranger]

Yes, please try the new release https://github.com/vllm-project/vllm-ascend/releases/tag/v0.7.1rc1

@wangxiyuan
This release does not work, and the problem remains unsolved. Only installing from the main branch works.

[wangxiyuan]

@dawnranger you should install a new torch-npu as well. Please follow the guide
https://vllm-ascend.readthedocs.io/en/v0.7.1rc1/installation.html#setup-vllm-and-vllm-ascend

[dawnranger]

@dawnranger you should install a new torch-npu as well. Please follow the guide https://vllm-ascend.readthedocs.io/en/v0.7.1rc1/installation.html#setup-vllm-and-vllm-ascend

@wangxiyuan torch_npu-2.5.1.dev20250218 has import error, see: #117