Closed weiji14 closed 2 days ago
/condalock Automatically locking new conda environment, building, and testing images...
Pulling in jaxlib
with the cuda120
build doesn't work yet, see https://github.com/pangeo-data/pangeo-docker-images/actions/runs/9181646795/job/25248830043#step:4:56:
INFO:conda_lock.conda_lock:Using virtual packages from virtual-packages.yml
Locking dependencies for ['linux-64']...
INFO:conda_lock.conda_solver:linux-64 using specs ['cuda-version >=12.0', 'flax >=0.8.0', 'jax', 'jaxlib >=0.4.23 cuda120*', 'jupyterlab-nvdashboard', 'keras-cv', 'tensorflow >=2.15.0 cuda120*', 'adlfs', 'argopy', 'awscli', 'black', 'boto3', 'bottleneck', 'cartopy', 'cdsapi', 'cfgrib', 'cf_xarray', 'ciso', 'cmocean', 'dask-ml', 'datashader', 'descartes', 'earthaccess', 'eofs', 'erddapy', 'esmpy', 'fastjmd95', 'flox', 'fsspec', 'gcm_filters', 'gcsfs', 'gh', 'gh-scoped-creds', 'geocube', 'geopandas', 'geopy', 'geoviews-core', 'git-lfs', 'gsw', 'h5netcdf', 'h5py', 'holoviews', 'hvplot', 'intake', 'intake-esm', 'intake-geopandas', 'intake-stac', 'intake-xarray', 'ipdb', 'ipykernel', 'ipyleaflet', 'ipytree', 'ipywidgets', 'jupyterlab_code_formatter', 'jupyterlab-git', 'jupyterlab-lsp', 'jupyterlab-myst', 'jupyter-panel-proxy', 'jupyter-resource-usage', 'kerchunk', 'line_profiler', 'lxml', 'lz4', 'matplotlib-base', 'memory_profiler', 'metpy', 'nb_conda_kernels', 'nbstripout', 'nc-time-axis', 'netcdf4', 'numbagg', 'numcodecs', 'numpy', 'numpy_groupies', 'odc-stac', 'pandas', 'panel', 'parcels', 'param', 'pop-tools', 'pyarrow', 'pycamhd', 'pydap', 'pystac', 'pystac-client', 'python-blosc', 'python-gist', 'python-graphviz', 'python-lsp-ruff', 'python-xxhash', 'rasterio', 'rechunker', 'rio-cogeo', 'rioxarray', 'ruff', 's3fs', 'satpy', 'scikit-image', 'scikit-learn', 'scipy', 'seaborn', 'sparse', 'snakeviz', 'stackstac', 'tiledb-py', 'timezonefinder', 'watermark', 'xarray', 'xarrayutils', 'xarray-datatree', 'xarray_leaflet', 'xarray-spatial', 'xbatcher', 'xcape', 'xclim', 'xesmf', 'xgboost', 'xgcm', 'xhistogram', 'xmip', 'xmitgcm', 'xpublish', 'xrft', 'xskillscore', 'xxhash', 'zarr', 'python 3.11.*', 'pangeo-notebook 2024.05.20.*', 'pip']
Failed to parse json, Expecting value: line 1 column 1 (char 0)
Could not lock the environment for platform linux-64
Could not solve for environment specs
The following packages are incompatible
├─ jaxlib >=0.4.23 cuda120* is installable with the potential options
│ ├─ jaxlib 0.4.23 would require
│ │ └─ libabseil >=20240116.1,<20240117.0a0 , which can be installed;
│ ├─ jaxlib 0.4.23 would require
│ │ └─ libabseil >=20240116.2,<20240117.0a0 , which can be installed;
│ └─ jaxlib 0.4.23 would require
│ └─ python_abi 3.12.* *_cp312, which requires
│ └─ python 3.12.* *_cpython, which can be installed;
├─ python 3.11** is not installable because it conflicts with any installable versions previously reported;
└─ tensorflow >=2.15.0 cuda120* is not installable because it requires
└─ tensorflow-base [2.15.0 cuda120py310heceb7ac_2|2.15.0 cuda120py310heceb7ac_3|...|2.15.0 cuda120py39hf42b710_3], which requires
└─ libabseil >=20230802.1,<20230803.0a0 , which conflicts with any installable versions previously reported.
{
"success": false
}
Need to wait for newer version of tensorflow
on conda-forge to use libabseil>=20240116
, wait for https://github.com/conda-forge/tensorflow-feedstock/pull/372 or https://github.com/conda-forge/tensorflow-feedstock/pull/385
Adding a explicit pin on
jaxlib=*=cuda120*
to pick up the cuda120 build version, ~as the conda-lock solver was picking up the cpu build of jaxlib instead of the cuda build in #514.~ See:https://github.com/pangeo-data/pangeo-docker-images/blob/6d4c2abcc5f145135e1d3213043b406fdf058793/ml-notebook/conda-lock.yml#L4166-L4182
Edit: the
cuda120
build is actually picked up automatically now as of the2024.06.02
tag, but setting thecuda120
pin still to be sure that this doesn't break in the future.Note that
jaxlib-0.4.23-cuda120py*
has an explicit runtime dependency oncuda-nvcc
since https://github.com/conda-forge/jaxlib-feedstock/pull/241, so this should mean users won't have to installcuda-nvcc
explicitly anymore.The
cuda-nvcc
workaround in the main README.md file has also been removed, in place of a message recommending users to useml-notebook>=2024.06.02
.Fixes #438