针对你遇到的ValueError: JAX requires ml_dtypes version 0.4.0 or newer; installed version问题,我们可以按照以下步骤进行解决: 检查当前安装的ml_dtypes版本: 你可以使用Python的包管理工具pip来检查当前安装的ml_dtypes版本。在命令行中运行以下命令: bash pip show ml_dtypes 这将显示ml_dtypes的详细信息,包...
dtypes.cc intn_numpy.h numpy.cc numpy.h ufuncs.h include tests __init__.py _finfo.py _iinfo.py py.typed third_party .clang-format .gitignore .gitmodules .pre-commit-config.yaml .pylintrc AUTHORS CHANGELOG.md CONTRIBUTING.md LICENSE ...
#if NPY_ABI_VERSION < 0x02000000 #define PyArray_DescrProto PyArray_Descr 2 changes: 1 addition & 1 deletion 2 ml_dtypes/_src/numpy.cc Original file line numberDiff line numberDiff line change @@ -17,7 +17,7 @@ limitations under the License. // ImportNumpy function to populate it....
To release a new version (e.g. from `1.0.0` -> `2.0.0`): * Create a new `# [2.0.0] - YYYY-MM-DD` header and add the current `[Unreleased]` notes. * At the end of the file: * Define the new link url: `[2.0.0]: https://github.com/jax-ml/ml_dtypes/compare/v1.0....
`ml_dtypes/__init__.py` to the correct version number, as well as updating `CHANGELOG.md` with the changes since the previous release (an example for the 0.2.0 release is [PR #78]). 1. Send a pull request updating the version in `ml_dtypes/__init__.py` to the new version num...
Try splitting wheel builds by python version. Build #358: Pull request #265 synchronize by copybara-service bot test_721980105 February 1, 2025 02:49 1m 1s Try splitting wheel builds by python version. Build #357: Pull request #265 opened by copybara-service bot test_721980105 Febr...
Language :: Python :: 3", "Programming Language :: Python :: 3 :: Only", "License :: OSI Approved :: Apache Software License", "Intended Audience :: Science/Research", ] keywords = [] # pip dependencies of the project dependencies = [ # Ensure numpy release supports Python version....
17 17 "__version__", 18 18 "bfloat16", 19 19 "finfo", 20 + "float8_e3m4", 20 21 "float8_e4m3", 21 22 "float8_e4m3b11fnuz", 22 23 "float8_e4m3fn", @@ -35,6 +36,7 @@ 35 36 from ml_dtypes._finfo import finfo 36 37 from ml_dtypes._iinfo import...
If you're exploring the use of low-precision dtypes in your code, you should be careful to anticipate when the precision loss might lead to surprising results. One example is the behavior of aggregations likesum; consider thisbfloat16summation in NumPy (run with version 1.24.2): ...
Version [e.g. 0.7.0] 0.11 Additional context Originally posted by @lkarthee in #1066 (comment) Activity lkartheechanged the title `np.ndarray` of bfloat16 using ml_dtypes is being interpreted as complex64 by mlx. `np.ndarray` of bfloat16 using ml_dtypes is being interpreted as complex...