Open maria-18-git opened 8 months ago
loadgen_wheel_recipe
on chai
:
axs byquery package_wheel,package_name=mlperf_loadgen
['^', 'byname', 'loadgen_package_wheel_for_python3.9']
Structure of created dicrectory:
maria@chai ~/work_collection/axs2mlperf/loadgen_wheel_recipe (master *=)$ tree -a `axs byquery package_wheel,package_name=mlperf_loadgen , get_path`
/home/maria/work_collection/loadgen_package_wheel_for_python3.9
├── data_axs.json
├── mlperf_loadgen-4.0-cp39-cp39-linux_x86_64.whl
└── mlperf_loadgen.egg-info
├── dependency_links.txt
├── PKG-INFO
├── SOURCES.txt
└── top_level.txt
1 directory, 6 files
axs byquery python_package,package_name=mlperf_loadgen,desired_python_version===3.9
...
"/home/maria/bin/python3" -m pip install /home/maria/work_collection/loadgen_package_wheel_for_python3.9/mlperf_loadgen-4.0-cp39-cp39-linux_x86_64.whl --prefix=/home/maria/work_collection/mlperf_loadgen_package_for_python3.9/install --ignore-installed
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
Looking in indexes: https://pypi.org/simple, https://pypi.ngc.nvidia.com
Processing /home/maria/work_collection/loadgen_package_wheel_for_python3.9/mlperf_loadgen-4.0-cp39-cp39-linux_x86_64.whl DEPRECATION: devscripts 2.22.1ubuntu1 has a non-standard version number. pip 24.0 will enforce this behaviour change. A possible replacement is to upgrade to a newer version of devscripts or contact the author to suggest that they release a version with a conforming version number. Discussion can be found at https://github.com/pypa/pip/issues/12063
Installing collected packages: mlperf-loadgen
Successfully installed mlperf-loadgen-4.0
[notice] A new release of pip is available: 23.3.2 -> 24.0
[notice] To update, run: python3 -m pip install --upgrade pip
INFO:root:Matched Rule #1/2 produced an entry, which matches the original query.
['^', 'byname', 'mlperf_loadgen_package_for_python3.9']
Structure of directory:
maria@chai ~/work_collection/axs2mlperf/loadgen_wheel_recipe (master *=)$ tree -a /home/maria/work_collection/mlperf_loadgen_package_for_python3.9/ /home/maria/work_collection/mlperf_loadgen_package_for_python3.9/
├── data_axs.json
└── install
└── lib
└── python3.9
└── site-packages
├── mlperf_loadgen-4.0.dist-info
│ ├── direct_url.json
│ ├── INSTALLER
│ ├── METADATA
│ ├── RECORD
│ ├── REQUESTED
│ ├── top_level.txt
│ └── WHEEL
└── mlperf_loadgen.cpython-39-x86_64-linux-gnu.so
5 directories, 9 files
mlperf_loadgen_package_for_python3.9
:maria@chai ~/work_collection/axs2mlperf/loadgen_wheel_recipe (master *=)$ axs byquery loadgen_output,task=bert,framework=onnxrt,loadgen_scenario=Offline,loadgen_mode=AccuracyOnly,execution_device=cpu
['^', 'byname', 'generated_by_bert_using_onnxrt_loadgen_on_get_14f91579f37040deafe3c11f1dcbdbca']
maria@chai ~/work_collection/axs2mlperf/loadgen_wheel_recipe (master *=)$ export ACCURACY_OUTPUT=`axs byquery loadgen_output,task=bert,framework=onnxrt,loadgen_scenario=Offline,loadgen_mode=AccuracyOnly,execution_device=cpu , get accuracy_dict`
maria@chai ~/work_collection/axs2mlperf/loadgen_wheel_recipe (master *=)$ echo "Accuracy: $ACCURACY_OUTPUT"
Accuracy: {'exact_match': 85.0, 'f1': 85.0}
maria@chai ~/work_collection/axs2mlperf/loadgen_wheel_recipe (master *=)$ axs byquery loadgen_output,task=object_detection,framework=onnxrt,loadgen_scenario=Offline,loadgen_mode=AccuracyOnly,model_name=ssd_resnet34,loadgen_dataset_size=20,loadgen_buffer_size=100,execution_device=cpu
['^', 'byname', 'generated_by_object_detection_using_onnxrt_loadgen_on_get_1ee92880af2a4eaa82f117dc1c19effe']
maria@chai ~/work_collection/axs2mlperf/loadgen_wheel_recipe (master *=)$ export ACCURACY_OUTPUT=`axs byquery loadgen_output,task=object_detection,framework=onnxrt,loadgen_scenario=Offline,loadgen_mode=AccuracyOnly,model_name=ssd_resnet34,loadgen_dataset_size=20,loadgen_buffer_size=100,execution_device=cpu , get accuracy`
maria@chai ~/work_collection/axs2mlperf/loadgen_wheel_recipe (master *=)$ echo "Accuracy: $ACCURACY_OUTPUT"
Accuracy: 22.852
maria@chai ~/work_collection/axs2mlperf/loadgen_wheel_recipe (master *=)$ echo `axs func round $ACCURACY_OUTPUT 0`
23.0
maria@chai ~/work_collection/axs2mlperf (master *=)$ axs byquery loadgen_output,task=object_detection,framework=onnxrt,loadgen_scenario=Offline,loadgen_mode=AccuracyOnly,model_name=retinanet_openimages,loadgen_dataset_size=20,loadgen_buffer_size=100,execution_device=cpu
['^', 'byname', 'generated_by_object_detection_using_onnxrt_loadgen_on_get_3e468e916d9c446e980202a79ede34af']
maria@chai ~/work_collection/axs2mlperf (master *=)$ export ACCURACY_OUTPUT=`axs byquery loadgen_output,task=object_detection,framework=onnxrt,loadgen_scenario=Offline,loadgen_mode=AccuracyOnly,model_name=retinanet_openimages,loadgen_dataset_size=20,loadgen_buffer_size=100,execution_device=cpu , get accuracy`
maria@chai ~/work_collection/axs2mlperf (master *=)$ echo "Accuracy: $ACCURACY_OUTPUT" Accuracy: 52.98
maria@chai ~/work_collection/axs2mlperf (master *=)$ axs byquery loadgen_output,task=image_classification,framework=onnxrt,loadgen_scenario=Offline,loadgen_mode=AccuracyOnly,model_name=resnet50,loadgen_dataset_size=20,loadgen_buffer_size=100,execution_device=cpu
['^', 'byname', 'generated_by_image_classification_using_onnxrt_loadgen_on_get_2f186ef2c94d44a8a92dfb544c7f04f1']
maria@chai ~/work_collection/axs2mlperf (master *=)$ export ACCURACY_OUTPUT=`axs byquery loadgen_output,task=image_classification,framework=onnxrt,loadgen_scenario=Offline,loadgen_mode=AccuracyOnly,model_name=resnet50,loadgen_dataset_size=20,loadgen_buffer_size=100,execution_device=cpu , get accuracy `
maria@chai ~/work_collection/axs2mlperf (master *=)$ echo "Accuracy: $ACCURACY_OUTPUT"
Accuracy: 85.0
maria@chai ~/work_collection/axs2mlperf (master *=)$ axs byquery loadgen_output,task=image_classification,framework=pytorch,loadgen_scenario=Offline,loadgen_mode=AccuracyOnly,model_name=resnet50,loadgen_dataset_size=20,loadgen_buffer_size=100,execution_device=cpu
['^', 'byname', 'generated_by_image_classification_using_torch_loadgen_on_get_2e326755c3b54c468b072d331639386d']
maria@chai ~/work_collection/axs2mlperf (master *=)$ export ACCURACY_OUTPUT=`axs byquery loadgen_output,task=image_classification,framework=pytorch,loadgen_scenario=Offline,loadgen_mode=AccuracyOnly,model_name=resnet50,loadgen_dataset_size=20,loadgen_buffer_size=100,execution_device=cpu , get accuracy`
maria@chai ~/work_collection/axs2mlperf (master *=)$ echo "Accuracy: $ACCURACY_OUTPUT" Accuracy: 75.0
mlperf_loadgen_package_for_python3.9
automaticallyaxs byquery package_wheel,package_name=mlperf_loadgen --- , remove
axs byquery python_package,package_name=mlperf_loadgen,desired_python_version===3.9 --- , remove
maria@chai ~/work_collection/axs2mlperf (master *=)$ axs byquery loadgen_output,task=bert,framework=onnxrt,loadgen_scenario=Offline,loadgen_mode=AccuracyOnly,execution_device=cpu
...
INFO:root:[work_collection] byquery(package_wheel,package_name=mlperf_loadgen,desired_python_version===3.9) did not find anything, but there are tags: {'package_wheel'} , trying to find a producer...
INFO:root:[work_collection] A total of 1 matched rules found.
INFO:root:Matched Rule #1/1: ['package_wheel', 'package_name=mlperf_loadgen'] from Entry 'loadgen_wheel_recipe'... INFO:root:Pipeline: [['get', 'pipeline']], Cumulative params: {'query': 'package_wheel,package_name=mlperf_loadgen,desired_python_version===3.9', 'package_name': 'mlperf_loadgen', 'desired_python_version': '3.9', 'tags': ['package_wheel'], 'cumulative_param_names': ['__query', 'package_name', 'desired_python_version', 'tags']} INFO:root:[work_collection] byquery(package_wheel,package_name=mlperf_loadgen,desired_python_version===3.9) did not find anything, but there are tags: {'package_wheel'} , trying to find a producer... INFO:root:[work_collection] A total of 1 matched rules found.
ERROR:root:RuntimeError: In pipeline [ ['get_kernel'] ['byquery', 'package_wheel,package_name=mlperf_loadgen,desired_python_version===3.9'] ['find_file', 'mlperf_loadgen.\.whl', {'return_full': True, 'abs_paths': True}] ] step ['find_file', 'mlperf_loadgen.\.whl', {'return_full': True, 'abs_paths': True}] cannot be executed on value (None) produced by ['byquery', 'package_wheel,package_name=mlperf_loadgen,desired_python_version===3.9'] None
axs byquery package_wheel,package_name=mlperf_loadgen,desired_python_version===3.9
doesn't work correctly with entry_creator
.
If
axs byquery package_wheel,package_name=mlperf_loadgen --- , remove
then
axs byquery python_package,package_name=mlperf_loadgen,desired_python_version===3.9
INFO:root:[work_collection] byquery(python_package,package_name=mlperf_loadgen,desired_python_version===3.9) did not find anything, but there are tags: {'python_package'} , trying to find a producer...
INFO:root:[work_collection] A total of 2 matched rules found.
INFO:root:Matched Rule #1/2: ['python_package', 'package_name=mlperf_loadgen', 'desired_python_version.'] from Entry 'loadgen_wheel_recipe'... INFO:root:Pipeline: [['get_kernel'], ['byname', 'pip'], ['install']], Cumulative params: {'__query': 'python_package,package_name=mlperf_loadgen,desired_python_version===3.9', 'installable': ['^^', 'execute', [[['get_kernel'], ['byquery', ['^^', 'substitute', 'package_wheel,package_name=mlperf_loadgen,desired_python_version===#{desired_python_version}#']], ['find_file', 'mlperf_loadgen.*\\.whl', {'return_full': True, 'abs_paths': True}]]]], 'package_name': 'mlperf_loadgen', 'desired_python_version': '3.9', 'tags': ['python_package'], '__cumulative_param_names': ['__query', 'installable', 'package_name', 'desired_python_version', 'tags']}
INFO:root:[work_collection] byquery(package_wheel,package_name=mlperf_loadgen,desired_python_version===3.9) did not find anything, but there are tags: {'package_wheel'} , trying to find a producer...
INFO:root:[work_collection] A total of 1 matched rules found.
INFO:root:Matched Rule #1/1: ['package_wheel', 'package_name=mlperf_loadgen'] from Entry 'loadgen_wheel_recipe'... INFO:root:Pipeline: [['get', 'pipeline']], Cumulative params: {'__query': 'package_wheel,package_name=mlperf_loadgen,desired_python_version===3.9', 'package_name': 'mlperf_loadgen', 'desired_python_version': '3.9', 'tags': ['package_wheel'], '__cumulative_param_names': ['__query', 'package_name', 'desired_python_version', 'tags']} INFO:root:[work_collection] byquery(package_wheel,package_name=mlperf_loadgen,desired_python_version===3.9) did not find anything, but there are tags: {'package_wheel'} , trying to find a producer...
INFO:root:[work_collection] A total of 1 matched rules found.
INFO:root:Matched Rule #1/1: ['package_wheel', 'package_name=mlperf_loadgen'] from Entry 'loadgen_wheel_recipe'... INFO:root:Pipeline: [['get', 'pipeline']], Cumulative params: {'__query': 'package_wheel,package_name=mlperf_loadgen,desired_python_version===3.9', 'package_name': 'mlperf_loadgen', 'desired_python_version': '3.9', 'tags': ['package_wheel'], '__cumulative_param_names': ['__query', 'package_name', 'desired_python_version', 'tags']} WARNING:root:[loadgen_wheel_recipe -> loadgen_wheel_recipe] parameter 'pipeline' is contained here, but BLOCKED by this entry -- all blockers: {'loadgen_wheel_recipe'} INFO:root:Matched Rule #1/1 didn't produce a result, 0 more matched rules to try...
------------------------------------------------------------------------------------------------------------------------ While computing nested_calls in ['^^', 'execute', [[['get_kernel'], ['byquery', ['^^', 'substitute', 'package_wheel,package_name=mlperf_loadgen,desired_python_version===#{desired_python_version}#']], ['find_file', 'mlperf_loadgen.*\\.whl', {'return_full': True, 'abs_paths': True}]]]] the following exception was raised: In pipeline [
['get_kernel'] ['byquery', 'package_wheel,package_name=mlperf_loadgen,desired_python_version===3.9']
['find_file', 'mlperf_loadgen.*\\.whl', {'return_full': True, 'abs_paths': True}]
] step ['find_file', 'mlperf_loadgen.*\\.whl', {'return_full': True, 'abs_paths': True}] cannot be executed on value (None) produced by ['byquery', 'package_wheel,package_name=mlperf_loadgen,desired_python_version===3.9']
...
While computing nested_calls in {'__query': 'python_package,package_name=mlperf_loadgen,desired_python_version===3.9', 'installable': ['^^', 'execute', [[['get_kernel'], ['byquery', ['^^', 'substitute', 'package_wheel,package_name=mlperf_loadgen,desired_python_version===#{desired_python_version}#']], ['find_file', 'mlperf_loadgen.*\\.whl', {'return_full': True, 'abs_paths': True}]]]], 'package_name': 'mlperf_loadgen', 'desired_python_version': '3.9', 'tags': ['python_package'], '__cumulative_param_names': ['__query', 'installable', 'package_name', 'desired_python_version', 'tags']} the following exception was raised: In pipeline [
['get_kernel']
['byquery', 'package_wheel,package_name=mlperf_loadgen,desired_python_version===3.9'] ['find_file', 'mlperf_loadgen.*\\.whl', {'return_full': True, 'abs_paths': True}]
] step ['find_file', 'mlperf_loadgen.*\\.whl', {'return_full': True, 'abs_paths': True}] cannot be executed on value (None) produced by ['byquery', 'package_wheel,package_name=mlperf_loadgen,desired_python_version===3.9']
======================================================================================================================== ERROR:root:RuntimeError: In pipeline [
['get_kernel']
['byquery', 'package_wheel,package_name=mlperf_loadgen,desired_python_version===3.9']
['find_file', 'mlperf_loadgen.*\\.whl', {'return_full': True, 'abs_paths': True}] ] step ['find_file', 'mlperf_loadgen.*\\.whl', {'return_full': True, 'abs_paths': True}] cannot be executed on value (None) produced by ['byquery', 'package_wheel,package_name=mlperf_loadgen,desired_python_version===3.9']
None
It is the same issue.
It is an issue with entry_creator
in loadgen_wheel_recipe/data_axs.json. We don't use entry_creator
many times.
All
axs
entries should useentry_creator
for object creation. Now different types ofsave
are used for it. For example,set_path
(entry_name setting) is used in case ofrecord_entry
.First entries without dependencies on others should be updated and then it should be tested. Also we need to check whether we can move
pipeline
parameter , e.g.to
entry_creator
or not.