[doc] fix doc build error caused by msgspec #7659
Conversation
|
👋 Hi! Thank you for contributing to the vLLM project. Once the PR is approved and ready to go, please make sure to run full CI as it is required to merge (or just use auto-merge). To run full CI, you can do one of these:
🚀 |
There was a problem hiding this comment.
LGTM if sufficient. It seems there is now a strange nvml error in doc build
[2024-08-19T18:38:40Z] Exception occurred:
--
| [2024-08-19T18:38:40Z] File "/usr/local/lib/python3.10/dist-packages/pynvml.py", line 979, in _nvmlCheckReturn
| [2024-08-19T18:38:40Z] raise NVMLError(ret)
| [2024-08-19T18:38:40Z] pynvml.NVMLError_LibraryNotFound: NVML Shared Library Not Found
| [2024-08-19T18:38:40Z] The full traceback has been saved in /tmp/sphinx-err-jlr90kw2.log, if you want to report the issue to the developers.
| [2024-08-19T18:38:40Z] Please also report this if it was a user error, so that a better error message can be provided next time.
| [2024-08-19T18:38:40Z] A bug report can be filed in the tracker at <https://github.com/sphinx-doc/sphinx/issues>. Thanks!
| [2024-08-19T18:38:40Z] make: *** [Makefile:20: html] Error 2
| [2024-08-19T18:38:41Z] 🚨 Error: The command exited with status 2
| # NOTE: we don't use `torch.version.cuda` / `torch.version.hip` because | ||
| # they only indicate the build configuration, not the runtime environment. | ||
| # For example, people can install a cuda build of pytorch but run on tpu. | ||
|
|
||
| is_tpu = False | ||
| try: | ||
| import torch_xla.core.xla_model as xm | ||
| xm.xla_device(devkind="TPU") | ||
| is_tpu = True | ||
| except Exception: | ||
| pass |
There was a problem hiding this comment.
cc @WoosukKwon
we don't use any successful import as a flag. Instead, only when some device code executes successfully, then we trust that we are in the current platform.
technically, we can install libtpu python package for any platform.
Signed-off-by: Alvant <alvasian@yandex.ru>
caused by #7109
adding
msgspecshould solve this.cc @simon-mo @khluu @ywang96 @DarkLight1337 , let's find a permanent solution.