Use legacy 1.21 for numpy - #155
Conversation
![chatgpt-codex-connector[bot]](https://avatars.githubusercontent.com/in/1144995?s=60&v=4){kind=small}
There was a problem hiding this comment.
💡 Codex Review
Here are some automated review suggestions for this pull request.
ℹ️ About Codex in GitHub
Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you
- Open a pull request for review
- Mark a draft as ready
- Comment "@codex review".
If Codex has suggestions, it will comment; otherwise it will react with 👍.
Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".
| os.environ['TF_CPP_MIN_LOG_LEVEL'] = '3' | ||
| os.environ['CUDA_VISIBLE_DEVICES'] = '' | ||
| np.set_printoptions(legacy="1.25") | ||
| np.set_printoptions(legacy="1.21") |
There was a problem hiding this comment.
The new call np.set_printoptions(legacy="1.21") is not a supported value for the legacy keyword—NumPy only accepts None, '1.13' and (since 2.0) '1.25'. With any released NumPy, this line raises ValueError: legacy must be None or '1.13'/'1.25' during module import, so the server fails before any logic runs. If the intent was to preserve the 1.25 print behaviour under NumPy 2.0, the previous '1.25' string was the correct option.
Useful? React with 👍 / 👎.
1 participant
No description provided.